{ "entries": [ { "company_id": "scale_ai", "products": [ { "name": "Scale Data Engine", "description": "Training data and annotation (labels added to data so models can learn).", "record_ids": [ "scale_ai.product.oct14-0", "scale_ai.product.major-0" ], "claim_ids": [ "scale_ai.claim.oct14-product-0", "scale_ai.claim.major-products-0" ] }, { "name": "Scale GenAI Portfolio", "description": "AI applications for work inside businesses.", "record_ids": [ "scale_ai.product.oct14-1", "scale_ai.product.major-0" ], "claim_ids": [ "scale_ai.claim.oct14-product-1", "scale_ai.claim.major-products-0" ] }, { "name": "Scale Donovan", "description": "AI applications for government users.", "record_ids": [ "scale_ai.product.oct14-2", "scale_ai.product.major-0" ], "claim_ids": [ "scale_ai.claim.oct14-product-2", "scale_ai.claim.major-products-0" ] } ], "relationships": [ { "name": "Defense Innovation Unit", "relationship": "customer", "description": "The buyer announced the Thunderforge prototype contract. Completed deployment is unverified.", "record_ids": [ "scale_ai.customer.major-0" ], "claim_ids": [ "scale_ai.claim.major-customers-0" ] }, { "name": "Toyota", "relationship": "historical_customer", "description": "Scale described this customer’s work in April 2021. Current status is unverified.", "record_ids": [ "scale_ai.customer.major-1" ], "claim_ids": [ "scale_ai.claim.major-customers-1" ] }, { "name": "States Title", "relationship": "historical_customer", "description": "Scale described this customer’s work in April 2021. Current status is unverified.", "record_ids": [ "scale_ai.customer.major-1" ], "claim_ids": [ "scale_ai.claim.major-customers-1" ] }, { "name": "PayPal", "relationship": "historical_customer", "description": "Scale described this customer’s work in April 2021. Current status is unverified.", "record_ids": [ "scale_ai.customer.major-1" ], "claim_ids": [ "scale_ai.claim.major-customers-1" ] }, { "name": "Brex", "relationship": "historical_customer", "description": "Scale described this customer’s work in April 2021. Current status is unverified.", "record_ids": [ "scale_ai.customer.major-1" ], "claim_ids": [ "scale_ai.claim.major-customers-1" ] }, { "name": "Flexport", "relationship": "historical_customer", "description": "Scale described this customer’s work in April 2021. Current status is unverified.", "record_ids": [ "scale_ai.customer.major-1" ], "claim_ids": [ "scale_ai.claim.major-customers-1" ] }, { "name": "Mayo Clinic", "relationship": "claimed_customer", "description": "Scale named this company in new business bookings for Q4 2025.", "record_ids": [ "scale_ai.customer.major-2" ], "claim_ids": [ "scale_ai.claim.major-customers-2" ] }, { "name": "BP", "relationship": "claimed_customer", "description": "Scale named this company in new business bookings for Q4 2025.", "record_ids": [ "scale_ai.customer.major-2" ], "claim_ids": [ "scale_ai.claim.major-customers-2" ] }, { "name": "Allianz", "relationship": "claimed_customer", "description": "Scale named this company in new business bookings for Q4 2025.", "record_ids": [ "scale_ai.customer.major-2" ], "claim_ids": [ "scale_ai.claim.major-customers-2" ] }, { "name": "Meta", "relationship": "strategic_investor", "description": "Strategic equity investor. Customer contract value is undisclosed.", "record_ids": [ "scale_ai.customer.major-3" ], "claim_ids": [ "scale_ai.claim.major-funding_rounds-6", "scale_ai.claim.major-customers-3" ] } ], "people": [ { "person_id": "scale_ai.person.alexandr-wang", "role": "Co-founder", "bio": "", "claim_ids": [ "scale_ai.claim.major-founders-0" ] }, { "person_id": "scale_ai.person.lucy-guo", "role": "Co-founder", "bio": "", "claim_ids": [ "scale_ai.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "surge_ai", "products": [ { "name": "RL Environments and Agents", "description": "Coding tasks for training and testing AI agents (software that takes actions to complete a task).", "record_ids": [ "surge_ai.product.oct14-0", "surge_ai.product.0", "surge_ai.product.1" ], "claim_ids": [ "surge_ai.claim.oct14-product-0", "surge_ai.claim.products-0", "surge_ai.claim.products-1" ] }, { "name": "Rubrics and Verifiers", "description": "Grading rules and checks for model outputs.", "record_ids": [ "surge_ai.product.oct14-1" ], "claim_ids": [ "surge_ai.claim.oct14-product-1" ] }, { "name": "RLHF", "description": "Human feedback for model training (people rate responses to guide model training).", "record_ids": [ "surge_ai.product.oct14-2" ], "claim_ids": [ "surge_ai.claim.oct14-product-2" ] }, { "name": "SFT", "description": "Supervised fine-tuning (models learn from examples of desired responses).", "record_ids": [ "surge_ai.product.oct14-3" ], "claim_ids": [ "surge_ai.claim.oct14-product-3" ] }, { "name": "Human Evaluation", "description": "Experts assess model responses and behavior.", "record_ids": [ "surge_ai.product.oct14-4", "surge_ai.product.0", "surge_ai.product.1" ], "claim_ids": [ "surge_ai.claim.oct14-product-4", "surge_ai.claim.products-0", "surge_ai.claim.products-1" ] }, { "name": "Off-The-Shelf Data", "description": "Ready-made datasets for model training.", "record_ids": [ "surge_ai.product.oct14-5", "surge_ai.product.0", "surge_ai.product.1" ], "claim_ids": [ "surge_ai.claim.oct14-product-5", "surge_ai.claim.products-0", "surge_ai.claim.products-1" ] }, { "name": "Post-Training Runs", "description": "Post-training (additional training to improve an existing model).", "record_ids": [ "surge_ai.product.oct14-6" ], "claim_ids": [ "surge_ai.claim.oct14-product-6" ] }, { "name": "EnterpriseBench: CoreCraft", "description": null, "record_ids": [ "surge_ai.product.oct14-7" ], "claim_ids": [ "surge_ai.claim.oct14-product-7" ] } ], "relationships": [ { "name": "OpenAI", "relationship": "historical_customer", "description": "OpenAI credits Surge for past WebGPT data collection. Current contract status and value are unverified.", "record_ids": [ "surge_ai.customer.0" ], "claim_ids": [ "surge_ai.claim.customers-0" ] }, { "name": "Anthropic", "relationship": "customer", "description": "Anthropic credits Surge with recruiting participants for persuasion research.", "record_ids": [ "surge_ai.customer.1" ], "claim_ids": [ "surge_ai.claim.customers-1" ] }, { "name": "Anthropic", "relationship": "claimed_partner", "description": "Surge lists Anthropic as a partner. Contract scope is undisclosed.", "record_ids": [ "surge_ai.customer.2" ], "claim_ids": [ "surge_ai.claim.customers-2" ] }, { "name": "Google", "relationship": "claimed_partner", "description": "Surge lists Google as a partner. Contract scope is undisclosed.", "record_ids": [ "surge_ai.customer.2" ], "claim_ids": [ "surge_ai.claim.customers-2" ] }, { "name": "Microsoft", "relationship": "claimed_partner", "description": "Surge lists Microsoft as a partner. Contract scope is undisclosed.", "record_ids": [ "surge_ai.customer.2" ], "claim_ids": [ "surge_ai.claim.customers-2" ] }, { "name": "Meta", "relationship": "claimed_partner", "description": "Surge lists Meta as a partner. Contract scope is undisclosed.", "record_ids": [ "surge_ai.customer.2" ], "claim_ids": [ "surge_ai.claim.customers-2" ] }, { "name": "Meta", "relationship": "customer", "description": "Meta papers name Surge as a provider of expert prompts and grading rules, plus a data-labeling platform. Contract value is undisclosed.", "record_ids": [ "surge_ai.customer.audit-0" ], "claim_ids": [ "surge_ai.claim.audit-meta-advancedif", "surge_ai.claim.audit-meta-mmrb2" ] } ], "people": [ { "person_id": "surge_ai.person.edwin-chen", "role": "Founder and CEO", "bio": "", "claim_ids": [ "surge_ai.claim.founders" ] } ], "notes": [] }, { "company_id": "mercor", "products": [ { "name": "Human data & evals", "description": "Expert-authored training data and evaluations.", "record_ids": [ "mercor.product.oct14-0", "mercor.product.major-0" ], "claim_ids": [ "mercor.claim.oct14-product-0", "mercor.claim.major-products-0" ] }, { "name": "APEX Benchmarks", "description": "Tests of professional work, with task datasets for AI agents.", "record_ids": [ "mercor.product.major-0" ], "claim_ids": [ "mercor.claim.major-products-0" ] }, { "name": "APEX-Agents", "description": "Tests extended professional work across several applications.", "record_ids": [ "mercor.product.round2-apex-agents" ], "claim_ids": [ "mercor.claim.round2-product-apex-agents" ] }, { "name": "APEX-Accounting", "description": "Tests accounting work across applications. Developed with Ramp.", "record_ids": [ "mercor.product.round2-apex-accounting" ], "claim_ids": [ "mercor.claim.round2-product-apex-accounting" ] }, { "name": "APEX-SWE", "description": "Tests software integration and observability (checking what software does while it runs). Developed with Cognition.", "record_ids": [ "mercor.product.round2-apex-swe" ], "claim_ids": [ "mercor.claim.round2-product-apex-swe" ] }, { "name": "Enterprise AI", "description": "Builds AI agents for business tasks.", "record_ids": [ "mercor.product.oct14-1", "mercor.product.major-0" ], "claim_ids": [ "mercor.claim.oct14-product-1", "mercor.claim.major-products-0" ] }, { "name": "Data monetization", "description": "Licenses data from businesses’ day-to-day operations.", "record_ids": [ "mercor.product.oct14-2", "mercor.product.major-0" ], "claim_ids": [ "mercor.claim.oct14-product-2", "mercor.claim.major-products-0" ] }, { "name": "Robotics", "description": "Robotics data for model training.", "record_ids": [ "mercor.product.major-0" ], "claim_ids": [ "mercor.claim.major-products-0" ] } ], "relationships": [ { "name": "OpenAI", "relationship": "reported_customer", "description": "Sacra reports an AI-data customer relationship. Current contracts and buyer confirmation are unverified.", "record_ids": [ "mercor.customer.major-0" ], "claim_ids": [ "mercor.claim.major-customers-0" ] }, { "name": "Anthropic", "relationship": "reported_customer", "description": "Sacra reports an AI-data customer relationship. Current contracts and buyer confirmation are unverified.", "record_ids": [ "mercor.customer.major-0" ], "claim_ids": [ "mercor.claim.major-customers-0" ] }, { "name": "Meta", "relationship": "reported_customer", "description": "Sacra reports an AI-data customer relationship. Current contracts and buyer confirmation are unverified.", "record_ids": [ "mercor.customer.major-0" ], "claim_ids": [ "mercor.claim.major-customers-0" ] } ], "people": [ { "person_id": "mercor.person.brendan-foody", "role": "Co-founder", "bio": "", "claim_ids": [ "mercor.claim.major-founders-0" ] }, { "person_id": "mercor.person.adarsh-hiremath", "role": "Co-founder", "bio": "", "claim_ids": [ "mercor.claim.major-founders-0" ] }, { "person_id": "mercor.person.surya-midha", "role": "Co-founder", "bio": "", "claim_ids": [ "mercor.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "handshake_ai", "products": [ { "name": "Handshake AI Fellowship", "description": "A network of professionals who create training data and test AI models.", "record_ids": [ "handshake_ai.product.oct14-0", "handshake_ai.product.major-0" ], "claim_ids": [ "handshake_ai.claim.oct14-product-0", "handshake_ai.claim.major-products-0" ] }, { "name": "Handshake AI Employer Data Partnerships", "description": null, "record_ids": [ "handshake_ai.product.oct14-1" ], "claim_ids": [ "handshake_ai.claim.oct14-product-1" ] }, { "name": "Enterprise AI", "description": null, "record_ids": [ "handshake_ai.product.oct14-2" ], "claim_ids": [ "handshake_ai.claim.oct14-product-2" ] } ], "relationships": [], "people": [ { "person_id": "handshake_ai.person.garrett-lord", "role": "Handshake co-founder", "bio": "", "claim_ids": [ "handshake_ai.claim.major-founders-0" ] }, { "person_id": "handshake_ai.person.ben-christensen", "role": "Handshake co-founder", "bio": "", "claim_ids": [ "handshake_ai.claim.major-founders-0" ] }, { "person_id": "handshake_ai.person.scott-ringwelski", "role": "Handshake co-founder", "bio": "", "claim_ids": [ "handshake_ai.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "micro1", "products": [ { "name": "Realm", "description": "Tests advanced AI models and provides reinforcement learning environments (models learn by trying tasks and receiving scores).", "record_ids": [ "micro1.product.major-0", "micro1.product.oct14-0" ], "claim_ids": [ "micro1.claim.major-products-0", "micro1.claim.oct14-product-0" ] }, { "name": "Cortex", "description": "Tests AI agents in live use and helps improve their performance.", "record_ids": [ "micro1.product.major-0", "micro1.product.oct14-1" ], "claim_ids": [ "micro1.claim.major-products-0", "micro1.claim.oct14-product-1" ] }, { "name": "Robotics", "description": "Training data from experts demonstrating real-world tasks.", "record_ids": [ "micro1.product.major-0", "micro1.product.oct14-2" ], "claim_ids": [ "micro1.claim.major-products-0", "micro1.claim.oct14-product-2" ] } ], "relationships": [ { "name": "Microsoft", "relationship": "claimed_customer", "description": "micro1 names this customer. The reviewed material does not say how it uses the service.", "record_ids": [ "micro1.customer.major-0" ], "claim_ids": [ "micro1.claim.major-customers-0" ] }, { "name": "Amazon", "relationship": "claimed_customer", "description": "Named as a customer in Forbes reporting.", "record_ids": [ "micro1.customer.major-1" ], "claim_ids": [ "micro1.claim.major-customers-1" ] }, { "name": "1X", "relationship": "claimed_customer", "description": "Named as a customer in Forbes reporting.", "record_ids": [ "micro1.customer.major-1" ], "claim_ids": [ "micro1.claim.major-customers-1" ] } ], "people": [ { "person_id": "micro1.person.ali-ansari", "role": "Founder", "bio": "", "claim_ids": [ "micro1.claim.major-founders-0" ] } ], "notes": [ { "text": "The October 12 company update values the company at $4 billion in its Series B (a funding round that helps a company expand). The amount of new funding and exact closing date remain undisclosed.", "claim_ids": [ "micro1.claim.oct14-series-b-priced" ] }, { "text": "Zara provides tools for recruiting and screening experts.", "claim_ids": [ "micro1.claim.major-products-0" ] } ] }, { "company_id": "turing", "products": [ { "name": "Frontier AI", "description": "Verified specialists produce coding and reasoning data for advanced AI models.", "record_ids": [ "turing.product.oct14-0", "turing.product.major-0" ], "claim_ids": [ "turing.claim.oct14-product-0", "turing.claim.major-products-0" ] }, { "name": "Off-The-Shelf Datasets", "description": "Ready-made datasets for model training.", "record_ids": [ "turing.product.oct14-1" ], "claim_ids": [ "turing.claim.oct14-product-1" ] }, { "name": "RL Environments", "description": "Training tasks with prompts and scoring tools.", "record_ids": [ "turing.product.oct14-2", "turing.product.major-0" ], "claim_ids": [ "turing.claim.oct14-product-2", "turing.claim.major-products-0" ] }, { "name": "Experts", "description": "Specialists who contribute to AI data work.", "record_ids": [ "turing.product.oct14-3" ], "claim_ids": [ "turing.claim.oct14-product-3" ] }, { "name": "Enterprise AI", "description": "Custom AI agents and deployment support for business tasks.", "record_ids": [ "turing.product.oct14-4", "turing.product.major-0" ], "claim_ids": [ "turing.claim.oct14-product-4", "turing.claim.major-products-0" ] } ], "relationships": [ { "name": "OpenAI", "relationship": "claimed_customer", "description": "Turing names this lab relationship. The buyer has not confirmed a paid AI-data contract in the reviewed evidence.", "record_ids": [ "turing.customer.major-0" ], "claim_ids": [ "turing.claim.major-customers-0" ] }, { "name": "Google", "relationship": "claimed_customer", "description": "Turing names this lab relationship. The buyer has not confirmed a paid AI-data contract in the reviewed evidence.", "record_ids": [ "turing.customer.major-0" ], "claim_ids": [ "turing.claim.major-customers-0" ] }, { "name": "Anthropic", "relationship": "claimed_customer", "description": "Turing names this lab relationship. The buyer has not confirmed a paid AI-data contract in the reviewed evidence.", "record_ids": [ "turing.customer.major-0" ], "claim_ids": [ "turing.claim.major-customers-0" ] }, { "name": "Coinbase", "relationship": "historical_customer", "description": "Named as a past talent-services customer. This evidence covers work outside AI data.", "record_ids": [ "turing.customer.major-1" ], "claim_ids": [ "turing.claim.major-customers-1" ] }, { "name": "Red Bull", "relationship": "historical_customer", "description": "Named as a past talent-services customer. This evidence covers work outside AI data.", "record_ids": [ "turing.customer.major-1" ], "claim_ids": [ "turing.claim.major-customers-1" ] }, { "name": "Reddit", "relationship": "historical_customer", "description": "Named as a past talent-services customer. This evidence covers work outside AI data.", "record_ids": [ "turing.customer.major-1" ], "claim_ids": [ "turing.claim.major-customers-1" ] }, { "name": "AWS", "relationship": "implementation_partner", "description": "Turing is an AWS Pattern Partner. AI-data purchasing by AWS is unverified.", "record_ids": [ "turing.customer.major-2" ], "claim_ids": [ "turing.claim.major-customers-2" ] }, { "name": "Anthropic", "relationship": "implementation_partner", "description": "Turing announced a Claude deployment partnership for healthcare and life sciences.", "record_ids": [ "turing.customer.major-3" ], "claim_ids": [ "turing.claim.major-customers-3" ] }, { "name": "HUMAIN", "relationship": "implementation_partner", "description": "Turing announced a marketplace for business AI agents. Financial terms are undisclosed.", "record_ids": [ "turing.customer.major-4" ], "claim_ids": [ "turing.claim.major-customers-4" ] } ], "people": [ { "person_id": "turing.person.jonathan-siddharth", "role": "Co-founder", "bio": "", "claim_ids": [ "turing.claim.major-founders-0" ] }, { "person_id": "turing.person.vijay-krishnan", "role": "Co-founder", "bio": "", "claim_ids": [ "turing.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "snorkel_ai", "products": [ { "name": "Snorkel Data Series", "description": "Specialist training data and reinforcement learning environments (models learn by trying tasks and receiving scores).", "record_ids": [ "snorkel_ai.product.oct14-0", "snorkel_ai.product.major-0" ], "claim_ids": [ "snorkel_ai.claim.oct14-product-0", "snorkel_ai.claim.major-products-0" ] }, { "name": "Agentic Data Platform", "description": "Combines expert work with AI quality checks.", "record_ids": [ "snorkel_ai.product.oct14-1", "snorkel_ai.product.major-0" ], "claim_ids": [ "snorkel_ai.claim.oct14-product-1", "snorkel_ai.claim.major-products-0" ] }, { "name": "Specialized agents", "description": "AI agents built for business applications.", "record_ids": [ "snorkel_ai.product.oct14-2", "snorkel_ai.product.major-0" ], "claim_ids": [ "snorkel_ai.claim.oct14-product-2", "snorkel_ai.claim.major-products-0" ] } ], "relationships": [ { "name": "BNY", "relationship": "claimed_customer", "description": "Snorkel announced a software and expertise agreement with a quote from BNY’s AI head.", "record_ids": [ "snorkel_ai.customer.major-0" ], "claim_ids": [ "snorkel_ai.claim.major-customers-0" ] }, { "name": "Experian", "relationship": "claimed_customer", "description": "Snorkel’s case study describes support emails handled with an adapted Llama model. Contract size is undisclosed.", "record_ids": [ "snorkel_ai.customer.major-1" ], "claim_ids": [ "snorkel_ai.claim.major-customers-1" ] }, { "name": "Wayfair", "relationship": "historical_customer", "description": "Named in past Snorkel releases. Current paid status is unverified.", "record_ids": [ "snorkel_ai.customer.major-2" ], "claim_ids": [ "snorkel_ai.claim.major-customers-2" ] }, { "name": "Chubb", "relationship": "historical_customer", "description": "Named in past Snorkel releases. Current paid status is unverified.", "record_ids": [ "snorkel_ai.customer.major-2" ], "claim_ids": [ "snorkel_ai.claim.major-customers-2" ] }, { "name": "Accenture", "relationship": "strategic_partner", "description": "Accenture confirmed investment and collaboration. Paid data purchasing is unverified.", "record_ids": [ "snorkel_ai.customer.major-3" ], "claim_ids": [ "snorkel_ai.claim.major-customers-3" ] } ], "people": [ { "person_id": "snorkel_ai.person.alex-ratner", "role": "Co-founder", "bio": "", "claim_ids": [ "snorkel_ai.claim.major-founders-0" ] }, { "person_id": "snorkel_ai.person.christopher-r", "role": "Co-founder", "bio": "", "claim_ids": [ "snorkel_ai.claim.major-founders-0" ] }, { "person_id": "snorkel_ai.person.paroma-varma", "role": "Co-founder", "bio": "", "claim_ids": [ "snorkel_ai.claim.major-founders-0" ] }, { "person_id": "snorkel_ai.person.braden-hancock", "role": "Co-founder", "bio": "", "claim_ids": [ "snorkel_ai.claim.major-founders-0" ] }, { "person_id": "snorkel_ai.person.henry-ehrenberg", "role": "Co-founder", "bio": "", "claim_ids": [ "snorkel_ai.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "labelbox", "products": [ { "name": "Horizon", "description": "Training environments where experts grade and evaluate model behavior.", "record_ids": [ "labelbox.product.oct14-1", "labelbox.product.major-0" ], "claim_ids": [ "labelbox.claim.oct14-product-1", "labelbox.claim.major-products-0" ] }, { "name": "WorldSim", "description": "Simulates business applications and data for agents training in Horizon.", "record_ids": [ "labelbox.product.round2-worldsim" ], "claim_ids": [ "labelbox.claim.round2-product-worldsim" ] }, { "name": "Recursion Managed Agents", "description": "AI agents for business tasks, with output scoring to improve specialist models.", "record_ids": [ "labelbox.product.oct14-0", "labelbox.product.major-0" ], "claim_ids": [ "labelbox.claim.oct14-product-0", "labelbox.claim.major-products-0" ] }, { "name": "Terra", "description": "Robotics datasets from first-person recordings and remotely controlled demonstrations.", "record_ids": [ "labelbox.product.oct14-2" ], "claim_ids": [ "labelbox.claim.oct14-product-2" ] }, { "name": "Alignerr", "description": "Qualified experts create preference data (ratings that show which responses people prefer). They also grade step-by-step task records.", "record_ids": [ "labelbox.product.oct14-3", "labelbox.product.major-0" ], "claim_ids": [ "labelbox.claim.oct14-product-3", "labelbox.claim.major-products-0" ] } ], "relationships": [ { "name": "Walmart", "relationship": "customer", "description": "Walmart engineering describes labeling for shopping and service AI. Use expanded beyond a pilot to more teams.", "record_ids": [ "labelbox.customer.major-0" ], "claim_ids": [ "labelbox.claim.major-customers-0" ] }, { "name": "Burberry", "relationship": "claimed_customer", "description": "Labelbox’s case study includes named Burberry comments about image labeling with Databricks.", "record_ids": [ "labelbox.customer.major-1" ], "claim_ids": [ "labelbox.claim.major-customers-1" ] }, { "name": "FLIR", "relationship": "claimed_customer", "description": "Labelbox’s Series C release quotes FLIR on operational use and planned expansion.", "record_ids": [ "labelbox.customer.major-2" ], "claim_ids": [ "labelbox.claim.major-customers-2" ] }, { "name": "Genentech", "relationship": "claimed_customer", "description": "Named in Labelbox’s Series C release. Buyer confirmation and 2026 contract status are unverified.", "record_ids": [ "labelbox.customer.major-3" ], "claim_ids": [ "labelbox.claim.major-customers-3" ] }, { "name": "Warner Brothers", "relationship": "claimed_customer", "description": "Named in Labelbox’s Series C release. Buyer confirmation and 2026 contract status are unverified.", "record_ids": [ "labelbox.customer.major-3" ], "claim_ids": [ "labelbox.claim.major-customers-3" ] }, { "name": "Bayer", "relationship": "claimed_customer", "description": "Named in Labelbox’s Series C release. Buyer confirmation and 2026 contract status are unverified.", "record_ids": [ "labelbox.customer.major-3" ], "claim_ids": [ "labelbox.claim.major-customers-3" ] }, { "name": "BASF", "relationship": "claimed_customer", "description": "Named in Labelbox’s Series C release. Buyer confirmation and 2026 contract status are unverified.", "record_ids": [ "labelbox.customer.major-3" ], "claim_ids": [ "labelbox.claim.major-customers-3" ] }, { "name": "Condé Nast", "relationship": "historical_customer", "description": "Named in Labelbox’s seed release. Current customer status is unverified.", "record_ids": [ "labelbox.customer.major-4" ], "claim_ids": [ "labelbox.claim.major-customers-4" ] }, { "name": "Lytx", "relationship": "historical_customer", "description": "Named in Labelbox’s seed release. Current customer status is unverified.", "record_ids": [ "labelbox.customer.major-4" ], "claim_ids": [ "labelbox.claim.major-customers-4" ] }, { "name": "Genius Sports", "relationship": "historical_customer", "description": "Named in Labelbox’s seed release. Current customer status is unverified.", "record_ids": [ "labelbox.customer.major-4" ], "claim_ids": [ "labelbox.claim.major-customers-4" ] } ], "people": [ { "person_id": "labelbox.person.manu-sharma", "role": "Co-founder", "bio": "", "claim_ids": [ "labelbox.claim.major-founders-0" ] }, { "person_id": "labelbox.person.ysiad-ferreiras", "role": "Co-founder", "bio": "", "claim_ids": [ "labelbox.claim.major-founders-0" ] }, { "person_id": "labelbox.person.daniel-rasmuson", "role": "Co-founder", "bio": "", "claim_ids": [ "labelbox.claim.major-founders-0" ] }, { "person_id": "labelbox.person.brian-rieger", "role": "Co-founder", "bio": "", "claim_ids": [ "labelbox.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "invisible", "products": [ { "name": "Neuron", "description": "Data systems for business AI applications.", "record_ids": [ "invisible.product.oct14-0", "invisible.product.major-0" ], "claim_ids": [ "invisible.claim.oct14-product-0", "invisible.claim.major-products-0" ] }, { "name": "Atomic", "description": "Systems for day-to-day business workflows.", "record_ids": [ "invisible.product.oct14-1", "invisible.product.major-0" ], "claim_ids": [ "invisible.claim.oct14-product-1", "invisible.claim.major-products-0" ] }, { "name": "Meridial", "description": "An expert network for validation work.", "record_ids": [ "invisible.product.oct14-2", "invisible.product.major-0" ], "claim_ids": [ "invisible.claim.oct14-product-2", "invisible.claim.major-products-0" ] }, { "name": "Synapse", "description": "Tools for labeling data and evaluating models.", "record_ids": [ "invisible.product.oct14-3", "invisible.product.major-0" ], "claim_ids": [ "invisible.claim.oct14-product-3", "invisible.claim.major-products-0" ] }, { "name": "Axon", "description": "AI agents that automate business operations.", "record_ids": [ "invisible.product.oct14-4", "invisible.product.major-0" ], "claim_ids": [ "invisible.claim.oct14-product-4", "invisible.claim.major-products-0" ] } ], "relationships": [ { "name": "Cohere", "relationship": "claimed_customer", "description": "Invisible’s case study includes named buyer comments about evaluating Command A. Work also covered expert labeling and tasks in several languages.", "record_ids": [ "invisible.customer.major-0" ], "claim_ids": [ "invisible.claim.major-customers-0" ] }, { "name": "Charlotte Hornets", "relationship": "claimed_customer", "description": "Buyer comments on Invisible’s site describe use for the 2025 draft and checks of computer-vision results.", "record_ids": [ "invisible.customer.major-1" ], "claim_ids": [ "invisible.claim.major-customers-1" ] }, { "name": "Microsoft", "relationship": "claimed_customer", "description": "Invisible’s financing release names Microsoft as a model builder using its services. Contract details are unverified.", "record_ids": [ "invisible.customer.major-2" ], "claim_ids": [ "invisible.claim.major-customers-2" ] }, { "name": "AWS", "relationship": "claimed_customer", "description": "Invisible’s financing release names AWS as a model builder using its services. Contract details are unverified.", "record_ids": [ "invisible.customer.major-2" ], "claim_ids": [ "invisible.claim.major-customers-2" ] }, { "name": "SwissGear", "relationship": "claimed_customer", "description": "Named enterprise customer example in Invisible’s financing release.", "record_ids": [ "invisible.customer.major-3" ], "claim_ids": [ "invisible.claim.major-customers-3" ] }, { "name": "SAIC", "relationship": "claimed_customer", "description": "Named enterprise customer example in Invisible’s financing release.", "record_ids": [ "invisible.customer.major-3" ], "claim_ids": [ "invisible.claim.major-customers-3" ] } ], "people": [ { "person_id": "invisible.person.francis-pedraza", "role": "Founder", "bio": "", "claim_ids": [ "invisible.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "appen", "products": [ { "name": "Frontier Alignment", "description": "Human data and evaluation for post-training (additional training to improve an existing model).", "record_ids": [ "appen.product.oct14-0", "appen.product.major-0" ], "claim_ids": [ "appen.claim.oct14-product-0", "appen.claim.major-products-0" ] }, { "name": "Agentic AI", "description": null, "record_ids": [ "appen.product.oct14-1" ], "claim_ids": [ "appen.claim.oct14-product-1" ] }, { "name": "Multimodal and Speech AI", "description": null, "record_ids": [ "appen.product.oct14-2" ], "claim_ids": [ "appen.claim.oct14-product-2" ] }, { "name": "Physical AI", "description": null, "record_ids": [ "appen.product.oct14-3" ], "claim_ids": [ "appen.claim.oct14-product-3" ] }, { "name": "Model Integrity", "description": null, "record_ids": [ "appen.product.oct14-4" ], "claim_ids": [ "appen.claim.oct14-product-4" ] }, { "name": "Research Services", "description": null, "record_ids": [ "appen.product.oct14-5" ], "claim_ids": [ "appen.claim.oct14-product-5" ] }, { "name": "CrowdGen", "description": "A platform where contributors collect and label data.", "record_ids": [ "appen.product.oct14-6", "appen.product.major-0" ], "claim_ids": [ "appen.claim.oct14-product-6", "appen.claim.major-products-0" ] } ], "relationships": [ { "name": "Cohere", "relationship": "claimed_customer", "description": "Appen’s case study describes adapting models with specialist data and ranking preferred responses. Buyer-authored confirmation was not reviewed.", "record_ids": [ "appen.customer.major-0" ], "claim_ids": [ "appen.claim.major-customers-0" ] }, { "name": "Airbus", "relationship": "claimed_customer", "description": "Appen’s case study includes buyer comments about an image-labeling pilot.", "record_ids": [ "appen.customer.major-1" ], "claim_ids": [ "appen.claim.major-customers-1" ] }, { "name": "Google", "relationship": "historical_customer", "description": "Appen disclosed termination in January 2024 and wind-down of all Google projects by March 19, 2024.", "record_ids": [ "appen.customer.major-2" ], "claim_ids": [ "appen.claim.major-customers-2" ] }, { "name": "AWS", "relationship": "implementation_partner", "description": "Cloud provider and sales partner through AWS Marketplace. Appen data purchasing by AWS is unverified.", "record_ids": [ "appen.customer.major-3" ], "claim_ids": [ "appen.claim.major-customers-3" ] } ], "people": [ { "person_id": "appen.person.julie-vonwiller", "role": "Founder", "bio": "Founded Appen in 1996.", "claim_ids": [ "appen.claim.major-founders-0" ] }, { "person_id": "appen.person.chris-vonwiller", "role": "Co-founder", "bio": "Joined Appen after its founding.", "claim_ids": [ "appen.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "taskus", "products": [ { "name": "AI & Data Operations", "description": "People collect and label data. Experts also write responses for model training.", "record_ids": [ "taskus.product.0", "taskus.product.oct14-0" ], "claim_ids": [ "taskus.claim.product-0", "taskus.claim.oct14-product-0" ] }, { "name": "AI Safety & Alignment", "description": "Red teaming (deliberate attacks that reveal model weaknesses) and human feedback support safer model behavior.", "record_ids": [ "taskus.product.1", "taskus.product.oct14-1", "taskus.product.oct14-2" ], "claim_ids": [ "taskus.claim.product-1", "taskus.claim.oct14-product-1", "taskus.claim.oct14-product-2" ] }, { "name": "Digital customer experience and Trust & Safety", "description": "Outsourced customer support, plus review of content and financial crime.", "record_ids": [ "taskus.product.2" ], "claim_ids": [ "taskus.claim.product-2" ] }, { "name": "TaskVerse", "description": "Gig workers collect and label data for projects.", "record_ids": [ "taskus.product.0", "taskus.product.oct14-8" ], "claim_ids": [ "taskus.claim.product-0", "taskus.claim.oct14-product-8" ] }, { "name": "Data Feedback & Evaluations", "description": "Experts review responses and design tests of model performance.", "record_ids": [ "taskus.product.oct14-2", "taskus.product.0", "taskus.product.1" ], "claim_ids": [ "taskus.claim.oct14-product-2", "taskus.claim.product-0", "taskus.claim.product-1" ] }, { "name": "TaskGPT", "description": null, "record_ids": [ "taskus.product.oct14-3" ], "claim_ids": [ "taskus.claim.oct14-product-3" ] }, { "name": "AssistAI", "description": null, "record_ids": [ "taskus.product.oct14-4" ], "claim_ids": [ "taskus.claim.oct14-product-4" ] }, { "name": "PromptAI", "description": null, "record_ids": [ "taskus.product.oct14-5" ], "claim_ids": [ "taskus.claim.oct14-product-5" ] }, { "name": "CallAI", "description": null, "record_ids": [ "taskus.product.oct14-6" ], "claim_ids": [ "taskus.claim.oct14-product-6" ] }, { "name": "AuthorAI", "description": null, "record_ids": [ "taskus.product.oct14-7" ], "claim_ids": [ "taskus.claim.oct14-product-7" ] } ], "relationships": [ { "name": "Meta", "relationship": "customer", "description": "A filing identifies Meta as a paying customer responsible for 26% of FY2025 group revenue. This spans multiple businesses.", "record_ids": [ "taskus.customer.0" ], "claim_ids": [ "taskus.claim.customer-0" ] }, { "name": "NVIDIA", "relationship": "technology_vendor", "description": "Provides technology supporting TaskGPT.", "record_ids": [ "taskus.customer.3" ], "claim_ids": [ "taskus.claim.customer-3" ] }, { "name": "Amazon Web Services", "relationship": "technology_vendor", "description": "Provides technology supporting TaskGPT.", "record_ids": [ "taskus.customer.3" ], "claim_ids": [ "taskus.claim.customer-3" ] } ], "people": [ { "person_id": "taskus.person.bryce-maddock", "role": "Co-founder, CEO and chairperson", "bio": "Previously worked at Rodman & Renshaw and co-founded MW Entertainment. Holds a BA from New York University.", "claim_ids": [ "taskus.claim.founder-0" ] }, { "person_id": "taskus.person.jaspar-weir", "role": "Co-founder and president", "bio": "Previously co-founded MW Entertainment. Holds a bachelor’s degree from the University of Southern California.", "claim_ids": [ "taskus.claim.founder-1" ] } ], "notes": [ { "text": "AI Services is one business line reported by TaskUs. Whole-company figures also include its broader outsourcing work.", "claim_ids": [ "taskus.claim.identity" ] }, { "text": "TaskUs remained publicly listed after its proposed 2025 buyout failed to win shareholder approval.", "claim_ids": [ "taskus.claim.ownership" ] } ] }, { "company_id": "imerit", "products": [ { "name": "Ango Hub", "description": "Software for labeling and reviewing data, including medical images and LiDAR (laser measurements used to map three-dimensional surroundings).", "record_ids": [ "imerit.product.imerit-ango", "imerit.product.oct14-0" ], "claim_ids": [ "imerit.claim.expansion-imerit-ango", "imerit.claim.oct14-product-0" ] }, { "name": "Deep Reasoning Lab", "description": "Experts design tasks with several steps. They use model errors to create corrective training data.", "record_ids": [ "imerit.product.imerit-deep-reasoning", "imerit.product.oct14-2" ], "claim_ids": [ "imerit.claim.expansion-imerit-deep-reasoning", "imerit.claim.oct14-product-2" ] }, { "name": "iMerit Scholars", "description": "Domain specialists create training data and evaluate model reasoning.", "record_ids": [ "imerit.product.imerit-scholars", "imerit.product.oct14-1" ], "claim_ids": [ "imerit.claim.expansion-imerit-scholars", "imerit.claim.oct14-product-1" ] }, { "name": "Specialized data services", "description": "Managed data labeling and evaluation for healthcare and robotics, including autonomous vehicles.", "record_ids": [ "imerit.product.imerit-services" ], "claim_ids": [ "imerit.claim.expansion-imerit-services" ] }, { "name": "3D Multi Sensor Fusion", "description": null, "record_ids": [ "imerit.product.oct14-3" ], "claim_ids": [ "imerit.claim.oct14-product-3" ] }, { "name": "Radiology Annotation Co-Pilot", "description": null, "record_ids": [ "imerit.product.oct14-4" ], "claim_ids": [ "imerit.claim.oct14-product-4" ] } ], "relationships": [ { "name": "Carbon Robotics", "relationship": "customer", "description": "Carbon confirms plant-image labeling work. iMerit’s case study reports roughly one million images. The team grew from five to over 80 specialists.", "record_ids": [ "imerit.customer.imerit-carbon", "imerit.customer.imerit-carbon-results" ], "claim_ids": [ "imerit.claim.expansion-imerit-carbon", "imerit.claim.expansion-imerit-carbon-results" ] } ], "people": [ { "person_id": "imerit.person.radha-ramaswami-basu", "role": "Founder; EXL Executive Vice President, Head of iMerit", "bio": "Spent 20 years at HP and later led Support.com through its IPO (the first sale of company shares to the public). Founded iMerit in 2012.", "claim_ids": [ "imerit.claim.expansion-imerit-founder", "imerit.claim.oct14-acquisition-completed" ] } ], "notes": [ { "text": "EXL completed its acquisition of iMerit on August 3, 2026.", "claim_ids": [ "imerit.claim.oct14-acquisition-completed" ] } ] }, { "company_id": "sama", "products": [ { "name": "Data annotation", "description": "A dedicated workforce labels images and video, as well as 3D data.", "record_ids": [ "sama.product.oct14-0", "sama.product.sama-annotate" ], "claim_ids": [ "sama.claim.oct14-product-0", "sama.claim.expansion-sama-annotate" ] }, { "name": "Data validation", "description": "Human review and correction of model predictions.", "record_ids": [ "sama.product.oct14-1", "sama.product.sama-validate" ], "claim_ids": [ "sama.claim.oct14-product-1", "sama.claim.expansion-sama-validate" ] }, { "name": "Model evaluation", "description": "Human evaluation of factual accuracy and instruction following.", "record_ids": [ "sama.product.oct14-2", "sama.product.sama-genai" ], "claim_ids": [ "sama.claim.oct14-product-2", "sama.claim.expansion-sama-genai" ] }, { "name": "Professional services", "description": null, "record_ids": [ "sama.product.oct14-3" ], "claim_ids": [ "sama.claim.oct14-product-3" ] }, { "name": "Sama Assure", "description": "A quality guarantee for delivered data.", "record_ids": [ "sama.product.oct14-4", "sama.product.sama-genai" ], "claim_ids": [ "sama.claim.oct14-product-4", "sama.claim.expansion-sama-genai" ] }, { "name": "Sama Hub", "description": "A shared workspace for data projects.", "record_ids": [ "sama.product.oct14-5", "sama.product.sama-genai" ], "claim_ids": [ "sama.claim.oct14-product-5", "sama.claim.expansion-sama-genai" ] }, { "name": "Sama IQ", "description": "Information about how data workflows perform.", "record_ids": [ "sama.product.oct14-6", "sama.product.sama-genai" ], "claim_ids": [ "sama.claim.oct14-product-6", "sama.claim.expansion-sama-genai" ] } ], "relationships": [ { "name": "Walmart", "relationship": "claimed_customer", "description": "Supplier case with named buyer testimony describes product-catalogue labeling and search relevance through 2019.", "record_ids": [ "sama.customer.sama-walmart" ], "claim_ids": [ "sama.claim.expansion-sama-walmart" ] }, { "name": "Google", "relationship": "claimed_customer", "description": "Supplier case with named buyer testimony describes polygon annotations for Project Guideline beginning July 2020.", "record_ids": [ "sama.customer.sama-google" ], "claim_ids": [ "sama.claim.expansion-sama-google" ] }, { "name": "Getty Images", "relationship": "claimed_customer", "description": "Supplier-hosted testimony from CTO Steve Heck discusses data volumes and quality.", "record_ids": [ "sama.customer.sama-testimonials" ], "claim_ids": [ "sama.claim.expansion-sama-testimonials" ] }, { "name": "Continental", "relationship": "claimed_customer", "description": "Supplier-hosted testimony from AI and Robotics Labs leader Demetrio Aiello discusses training-data strategy.", "record_ids": [ "sama.customer.sama-testimonials" ], "claim_ids": [ "sama.claim.expansion-sama-testimonials" ] }, { "name": "Microsoft", "relationship": "displayed_logo", "description": "Sama displays this logo. Separate evidence of contract scope and payment is unavailable.", "record_ids": [ "sama.customer.sama-logo-evidence" ], "claim_ids": [ "sama.claim.expansion-sama-logo-evidence" ] }, { "name": "eBay", "relationship": "displayed_logo", "description": "Sama displays this logo. Separate evidence of contract scope and payment is unavailable.", "record_ids": [ "sama.customer.sama-logo-evidence" ], "claim_ids": [ "sama.claim.expansion-sama-logo-evidence" ] }, { "name": "NASA", "relationship": "displayed_logo", "description": "Sama displays this logo. Separate evidence of contract scope and payment is unavailable.", "record_ids": [ "sama.customer.sama-logo-evidence" ], "claim_ids": [ "sama.claim.expansion-sama-logo-evidence" ] }, { "name": "Sony", "relationship": "displayed_logo", "description": "Sama displays this logo. Separate evidence of contract scope and payment is unavailable.", "record_ids": [ "sama.customer.sama-logo-evidence" ], "claim_ids": [ "sama.claim.expansion-sama-logo-evidence" ] }, { "name": "Qualcomm", "relationship": "displayed_logo", "description": "Sama displays this logo. Separate evidence of contract scope and payment is unavailable.", "record_ids": [ "sama.customer.sama-logo-evidence" ], "claim_ids": [ "sama.claim.expansion-sama-logo-evidence" ] }, { "name": "Verizon", "relationship": "displayed_logo", "description": "Sama displays this logo. Separate evidence of contract scope and payment is unavailable.", "record_ids": [ "sama.customer.sama-logo-evidence" ], "claim_ids": [ "sama.claim.expansion-sama-logo-evidence" ] }, { "name": "Siemens", "relationship": "displayed_logo", "description": "Sama displays this logo. Separate evidence of contract scope and payment is unavailable.", "record_ids": [ "sama.customer.sama-logo-evidence" ], "claim_ids": [ "sama.claim.expansion-sama-logo-evidence" ] }, { "name": "OpenAI", "relationship": "historical_customer", "description": "Independent reporting with buyer and supplier confirmation documents safety-labeling work ending in March 2022.", "record_ids": [ "sama.customer.sama-openai-historical" ], "claim_ids": [ "sama.claim.expansion-sama-openai-historical" ] }, { "name": "Meta", "relationship": "historical_customer", "description": "Independent reporting with buyer confirmation documents termination of the Nairobi AI-data labeling engagement in April 2026.", "record_ids": [ "sama.customer.sama-meta-termination" ], "claim_ids": [ "sama.claim.expansion-sama-meta-termination" ] } ], "people": [ { "person_id": "sama.person.leila-janah", "role": "Founder (died 2020)", "bio": "Founded Samasource in 2008 to expand paid digital work in underserved communities.", "claim_ids": [ "sama.claim.expansion-sama-founder" ] }, { "person_id": "sama.person.wendy-gonzalez", "role": "CEO", "bio": "Joined in 2015 and became CEO in 2020.", "claim_ids": [ "sama.claim.expansion-sama-current-ceo" ] } ], "notes": [] }, { "company_id": "abundant_ai", "products": [ { "name": "Custom RL environments and datasets", "description": "Simulation and model-training work for coding and physical science. Also covers recursive self-improvement (models working to improve their own capabilities) and alignment (guiding models toward intended and safe behavior).", "record_ids": [ "abundant_ai.product.oct14-0" ], "claim_ids": [ "abundant_ai.claim.oct14-product-0" ], "kind": "Commercial service" }, { "name": "SWE-gen", "description": "Turns code changes into verified coding tasks. Rebuilds a buggy starting point from working code to reduce errors caused by broken tests.", "record_ids": [ "abundant_ai.product.oct14-1" ], "claim_ids": [ "abundant_ai.claim.oct14-product-1" ], "kind": "Open-source environment generation" }, { "name": "SWE-Marathon", "description": "Tests extended software work, including rebuilding libraries and complete applications. Also covers machine-learning engineering.", "record_ids": [ "abundant_ai.product.oct14-2" ], "claim_ids": [ "abundant_ai.claim.oct14-product-2" ], "kind": "Research benchmark" }, { "name": "TheoremSmith", "description": "Creates difficult proof tasks in Lean (software that checks whether mathematical proofs are correct). Uses isolated checks to detect attempts to game rewards.", "record_ids": [ "abundant_ai.product.oct14-3" ], "claim_ids": [ "abundant_ai.claim.oct14-product-3" ], "kind": "Open-source research pipeline" } ], "relationships": [], "people": [ { "person_id": "abundant_ai.person.jesse-hu", "role": "Co-founder", "bio": "", "claim_ids": [ "abundant_ai.claim.founders" ] }, { "person_id": "abundant_ai.person.meji-abidoye", "role": "Co-founder", "bio": "", "claim_ids": [ "abundant_ai.claim.founders" ] }, { "person_id": "abundant_ai.person.ke-huang", "role": "Co-founder", "bio": "", "claim_ids": [ "abundant_ai.claim.founders" ] } ], "notes": [] }, { "company_id": "bespoke_labs", "products": [ { "name": "RL Environments & Infrastructure", "description": "Simulates company systems using real codebases and small connected services. Provides tools to build, run and improve these environments.", "record_ids": [ "bespoke_labs.product.oct14-0" ], "claim_ids": [ "bespoke_labs.claim.oct14-product-0" ], "kind": "Commercial platform/service" }, { "name": "GEPA", "description": "Uses task feedback to improve prompts and model behavior.", "record_ids": [ "bespoke_labs.product.oct14-1" ], "claim_ids": [ "bespoke_labs.claim.oct14-product-1" ], "kind": "Open research/optimizer" }, { "name": "Curator", "description": "Prepares synthetic data (examples generated by software rather than collected directly) for further training and structured extraction.", "record_ids": [ "bespoke_labs.product.oct14-2" ], "claim_ids": [ "bespoke_labs.claim.oct14-product-2" ], "kind": "Open-source library" }, { "name": "OpenThoughts", "description": "Open research on preparing data for reasoning models.", "record_ids": [ "bespoke_labs.product.oct14-3" ], "claim_ids": [ "bespoke_labs.claim.oct14-product-3" ], "kind": "Open reasoning datasets" }, { "name": "Terminal-Bench", "description": "Contributes tests of agent work in command-line environments (text-based interfaces for running commands on computers).", "record_ids": [ "bespoke_labs.product.oct14-4" ], "claim_ids": [ "bespoke_labs.claim.oct14-product-4" ], "kind": "Research contribution" }, { "name": "Nimble", "description": "Returns probabilities for allowed answers. Nimble-v4 and Nimble-v4-max support grading and agent decisions, including routing tasks.", "record_ids": [ "bespoke_labs.product.oct14-5" ], "claim_ids": [ "bespoke_labs.claim.oct14-product-5" ], "kind": "Hosted decision-model API" } ], "relationships": [ { "name": "Intuit Credit Karma", "relationship": "research_collaborator", "description": "Bespoke reports further training for SeeWhy and results from offline tests. Live deployment and payment are unverified.", "record_ids": [ "bespoke_labs.customer.0" ], "claim_ids": [ "bespoke_labs.claim.audit-seewhy" ] }, { "name": "Intuit", "relationship": "reported_customer", "description": "Silicon Valley Business Journal names Intuit as a customer.", "record_ids": [ "bespoke_labs.customer.1" ], "claim_ids": [ "bespoke_labs.claim.customers-1" ] }, { "name": "Meta", "relationship": "dataset_user", "description": "Bespoke’s blog reports use of OpenThoughts datasets. Purchased services are unverified.", "record_ids": [ "bespoke_labs.customer.2" ], "claim_ids": [ "bespoke_labs.claim.audit-dataset-relationships" ] }, { "name": "Amazon", "relationship": "dataset_user", "description": "Bespoke’s blog reports use of OpenThoughts datasets. Purchased services are unverified.", "record_ids": [ "bespoke_labs.customer.2" ], "claim_ids": [ "bespoke_labs.claim.audit-dataset-relationships" ] }, { "name": "AI2", "relationship": "dataset_user", "description": "Bespoke’s blog reports use of OpenThoughts datasets. Purchased services are unverified.", "record_ids": [ "bespoke_labs.customer.2" ], "claim_ids": [ "bespoke_labs.claim.audit-dataset-relationships" ] }, { "name": "Thinking Machines Lab", "relationship": "dataset_citer", "description": "Bespoke’s blog names citations of OpenThoughts. Purchased services are unverified.", "record_ids": [ "bespoke_labs.customer.2" ], "claim_ids": [ "bespoke_labs.claim.audit-dataset-relationships" ] }, { "name": "Microsoft", "relationship": "dataset_citer", "description": "Bespoke’s blog names citations of OpenThoughts. Purchased services are unverified.", "record_ids": [ "bespoke_labs.customer.2" ], "claim_ids": [ "bespoke_labs.claim.audit-dataset-relationships" ] }, { "name": "NVIDIA", "relationship": "dataset_citer", "description": "Bespoke’s blog names citations of OpenThoughts. Purchased services are unverified.", "record_ids": [ "bespoke_labs.customer.2" ], "claim_ids": [ "bespoke_labs.claim.audit-dataset-relationships" ] }, { "name": "Snowflake", "relationship": "research_collaborator", "description": "Snowflake AI Research confirms joint data-eng-bench work. Payment is undisclosed.", "record_ids": [ "bespoke_labs.customer.audit-snowflake" ], "claim_ids": [ "bespoke_labs.claim.audit-snowflake-partnership" ] } ], "people": [ { "person_id": "bespoke_labs.person.mahesh-sathiamoorthy", "role": "Co-founder", "bio": "Previously a staff engineer at Google DeepMind/Brain. Worked on systems for YouTube recommendations.", "claim_ids": [ "bespoke_labs.claim.audit-mahesh-sathiamoorthy-background" ] }, { "person_id": "bespoke_labs.person.alex-dimakis", "role": "Co-founder", "bio": "UC Berkeley electrical engineering and computer science professor and IEEE Fellow. Researches generative AI and information theory.", "claim_ids": [ "bespoke_labs.claim.audit-alex-dimakis-background" ] } ], "notes": [] }, { "company_id": "datacurve", "products": [ { "name": "RL environments", "description": "Extended tasks with natural instructions and realistic tools. Measures partial progress across software and data science, plus cybersecurity, machine learning and research.", "record_ids": [ "datacurve.product.oct14-0" ], "claim_ids": [ "datacurve.claim.oct14-product-0" ], "kind": "Commercial service" }, { "name": "OTS Datasets", "description": "Ready-made, reviewed datasets. Custom work includes supervised fine-tuning (models learn from examples of desired responses), expert task records and evaluations.", "record_ids": [ "datacurve.product.oct14-1" ], "claim_ids": [ "datacurve.claim.oct14-product-1" ], "kind": "Dataset catalog" }, { "name": "DeepSWE", "description": "Original engineering tasks written from scratch. Handwritten checks assess behavior; tasks are not adapted from existing code changes.", "record_ids": [ "datacurve.product.oct14-2" ], "claim_ids": [ "datacurve.claim.oct14-product-2" ], "kind": "Public benchmark" }, { "name": "Shipd", "description": "Pays for approved technical challenge submissions. Mars covers software engineering; Eris covers machine learning. Monte covers data science.", "record_ids": [ "datacurve.product.oct14-3" ], "claim_ids": [ "datacurve.claim.oct14-product-3" ], "kind": "Contributor marketplace" } ], "relationships": [], "people": [ { "person_id": "datacurve.person.serena-ge", "role": "Co-founder", "bio": "Studied computer science at Waterloo. Worked on synthetic-data reasoning and coding at Cohere.", "claim_ids": [ "datacurve.claim.audit-serena-ge-background" ] }, { "person_id": "datacurve.person.charley-lee", "role": "Co-founder", "bio": "Studied computer science at Waterloo. Researched reinforcement learning across data formats and browser agents.", "claim_ids": [ "datacurve.claim.audit-charley-lee-background" ] } ], "notes": [] }, { "company_id": "ooak", "products": [ { "name": "Enterprise digital twins", "description": "Removes identifying details from connected company communications, documents and tools. Preserves their relationships and the company’s structure.", "record_ids": [ "ooak.product.oct14-0" ], "claim_ids": [ "ooak.claim.oct14-product-0" ], "kind": "Custom data infrastructure" }, { "name": "RL environments", "description": "Tasks with several steps, based on real company work and calibrated against advanced models.", "record_ids": [ "ooak.product.oct14-1" ], "claim_ids": [ "ooak.claim.oct14-product-1" ], "kind": "Custom environments" }, { "name": "Company data licensing", "description": "Acquires connected business datasets. Handles extraction and delivery while removing identifying details across sources.", "record_ids": [ "ooak.product.oct14-2" ], "claim_ids": [ "ooak.claim.oct14-product-2" ], "kind": "Data-supply partnership" }, { "name": "Data Hub", "description": "Instructions for connecting and exporting company data. Named in the official first product-manager job description.", "record_ids": [ "ooak.product.oct14-3" ], "claim_ids": [ "ooak.claim.oct14-product-3" ], "kind": "Partner onboarding surface" } ], "relationships": [], "people": [ { "person_id": "ooak.person.pierre-louis-vouteau", "role": "Co-founder and CEO", "bio": "", "claim_ids": [ "ooak.claim.founders-0" ] }, { "person_id": "ooak.person.thomas-aubry", "role": "Co-founder and CTO", "bio": "", "claim_ids": [ "ooak.claim.founders-0" ] }, { "person_id": "ooak.person.gr-goire-lamy", "role": "Co-founder and COO", "bio": "", "claim_ids": [ "ooak.claim.founders-0" ] } ], "notes": [] }, { "company_id": "pareto", "products": [ { "name": "Expert training data", "description": "Captures specialist judgment in medicine and law. Also covers STEM, finance, software and the humanities.", "record_ids": [ "pareto.product.oct14-0" ], "claim_ids": [ "pareto.claim.oct14-product-0" ] }, { "name": "Forte", "description": "Experts register and are matched to projects. Matching is reviewed on an ongoing basis.", "record_ids": [ "pareto.product.oct14-1" ], "claim_ids": [ "pareto.claim.oct14-product-1" ] }, { "name": "LEAPBench", "description": "Measures how efficiently models learn across repeated scientific design tasks. Also assesses their final results.", "record_ids": [ "pareto.product.oct14-2" ], "claim_ids": [ "pareto.claim.oct14-product-2" ] }, { "name": "AttuneBench", "description": "Tests conversational emotional intelligence using participants’ own ratings.", "record_ids": [ "pareto.product.oct14-3" ], "claim_ids": [ "pareto.claim.oct14-product-3" ] } ], "relationships": [ { "name": "Anthropic", "relationship": "claimed_customer", "description": "An April 4, 2026 employer posting claims work in live use. Current status and completed renewals are unconfirmed.", "record_ids": [ "pareto.customer.0" ], "claim_ids": [ "pareto.claim.customers-0" ] }, { "name": "Google DeepMind", "relationship": "claimed_customer", "description": "An April 4, 2026 employer posting claims work in live use. Current status and completed renewals are unconfirmed.", "record_ids": [ "pareto.customer.0" ], "claim_ids": [ "pareto.claim.customers-0" ] }, { "name": "Character.AI", "relationship": "research_collaborator", "description": "Older employer materials describe research collaboration. Payment is unverified.", "record_ids": [ "pareto.customer.1" ], "claim_ids": [ "pareto.claim.customers-1", "pareto.claim.audit-pareto-collaborator-tag-pareto.customer.1" ] }, { "name": "Imbue", "relationship": "research_collaborator", "description": "Older employer materials describe research collaboration. Payment is unverified.", "record_ids": [ "pareto.customer.1" ], "claim_ids": [ "pareto.claim.customers-1", "pareto.claim.audit-pareto-collaborator-tag-pareto.customer.1" ] }, { "name": "Stanford", "relationship": "research_collaborator", "description": "Older employer materials describe research collaboration. Payment is unverified.", "record_ids": [ "pareto.customer.1" ], "claim_ids": [ "pareto.claim.customers-1", "pareto.claim.audit-pareto-collaborator-tag-pareto.customer.1" ] }, { "name": "UPenn", "relationship": "research_collaborator", "description": "Older employer materials describe research collaboration. Payment is unverified.", "record_ids": [ "pareto.customer.1" ], "claim_ids": [ "pareto.claim.customers-1", "pareto.claim.audit-pareto-collaborator-tag-pareto.customer.1" ] }, { "name": "UK AI Security Institute", "relationship": "customer", "description": "The buyer confirms expert recruitment and grading rules for evaluating harm. Work also included training data to detect harmful advice.", "record_ids": [ "pareto.customer.uk-aisi", "pareto.customer.audit-pareto-aisi-outcome" ], "claim_ids": [ "pareto.claim.audit-pareto-aisi-customer", "pareto.claim.audit-pareto-aisi-outcome" ] } ], "people": [ { "person_id": "pareto.person.phoebe-yao", "role": "Founder and CEO", "bio": "", "claim_ids": [ "pareto.claim.founders-0" ] } ], "notes": [] }, { "company_id": "polymath_ai_labs", "products": [ { "name": "Simulation environments", "description": "Simulated working systems whose data and tools change over time. Outcomes can be checked.", "record_ids": [ "polymath_ai_labs.product.oct14-0" ], "claim_ids": [ "polymath_ai_labs.claim.oct14-product-0" ] }, { "name": "Horizon-SWE", "description": "Tests software work from start to finish inside a simulated company, including deployment and monitoring.", "record_ids": [ "polymath_ai_labs.product.oct14-1" ], "claim_ids": [ "polymath_ai_labs.claim.oct14-product-1" ] } ], "relationships": [], "people": [ { "person_id": "polymath_ai_labs.person.dylan-ma", "role": "Co-founder", "bio": "Previously led further model-training research and data at Hume AI. Studied computer science at Berkeley.", "claim_ids": [ "polymath_ai_labs.claim.audit-dylan-ma-background" ] }, { "person_id": "polymath_ai_labs.person.naren-yenuganti", "role": "Co-founder", "bio": "Built credit-monitoring systems at Plaid. Worked on machine-learning logistics systems at Amazon.", "claim_ids": [ "polymath_ai_labs.claim.audit-naren-yenuganti-background" ] } ], "notes": [] }, { "company_id": "proximal", "products": [ { "name": "Data research and post-training infrastructure", "description": "Turns work records and outputs into tests of model abilities and targeted training data.", "record_ids": [ "proximal.product.oct14-0" ], "claim_ids": [ "proximal.claim.oct14-product-0" ] }, { "name": "FrontierSWE v2", "description": "Tests extended technical work in implementation and performance. Tasks also cover science, visual reasoning and AI research.", "record_ids": [ "proximal.product.oct14-1" ], "claim_ids": [ "proximal.claim.oct14-product-1" ] }, { "name": "Proximus", "description": "A testing system designed to keep agents working within a 20-hour task budget.", "record_ids": [ "proximal.product.oct14-2" ], "claim_ids": [ "proximal.claim.oct14-product-2" ] } ], "relationships": [ { "name": "Modular", "relationship": "research_collaborator", "description": "FrontierSWE acknowledgements name this research partner. A buying relationship is unverified.", "record_ids": [], "claim_ids": [ "proximal.claim.audit-research-partners" ] }, { "name": "Prime Intellect", "relationship": "research_collaborator", "description": "FrontierSWE acknowledgements name this research partner. A buying relationship is unverified.", "record_ids": [], "claim_ids": [ "proximal.claim.audit-research-partners" ] }, { "name": "Thoughtful Lab", "relationship": "research_collaborator", "description": "FrontierSWE acknowledgements name this research partner. A buying relationship is unverified.", "record_ids": [], "claim_ids": [ "proximal.claim.audit-research-partners" ] } ], "people": [ { "person_id": "proximal.person.calvin-chen", "role": "Co-founder", "bio": "", "claim_ids": [ "proximal.claim.audit-calvin-chen-background" ] }, { "person_id": "proximal.person.justus-mattern", "role": "Co-founder", "bio": "", "claim_ids": [ "proximal.claim.audit-justus-mattern-background" ] }, { "person_id": "proximal.person.navid-pour", "role": "Co-founder", "bio": "", "claim_ids": [ "proximal.claim.audit-third-founder", "proximal.claim.audit-navid-background" ] } ], "notes": [] }, { "company_id": "quesma", "products": [ { "name": "Quesma coding-agent analytics", "description": "Analyzes coding-agent sessions and costs. The reviewed product has an early-access waitlist.", "record_ids": [ "quesma.product.0", "quesma.product.2", "quesma.product.oct14-0" ], "claim_ids": [ "quesma.claim.products-0", "quesma.claim.products-2", "quesma.claim.oct14-product-0" ] }, { "name": "Quesma Shipper", "description": "An open-source collector that saves agent-session transcripts in the user’s Amazon S3 storage.", "record_ids": [ "quesma.product.oct14-1" ], "claim_ids": [ "quesma.claim.oct14-product-1" ] }, { "name": "CompileBench / OTelBench / BinaryAudit / BabaIsBench", "description": "These four benchmarks (shared tests for comparing model performance) cover specific engineering or puzzle tasks.", "record_ids": [ "quesma.product.1", "quesma.product.oct14-2" ], "claim_ids": [ "quesma.claim.products-1", "quesma.claim.oct14-product-2" ] } ], "relationships": [ { "name": "Hydrolix", "relationship": "former_customer", "description": "Hydrolix confirms a past reseller relationship. It also bought the intellectual property for Quesma’s database gateway.", "record_ids": [ "quesma.customer.0" ], "claim_ids": [ "quesma.claim.audit-historical-delivery" ] }, { "name": "OpenAI", "relationship": "research_citation", "description": "OpenAI cited Quesma authors’ Baba Is You research.", "record_ids": [ "quesma.customer.2" ], "claim_ids": [ "quesma.claim.audit-relationship-0" ] } ], "people": [ { "person_id": "quesma.person.jacek-migda", "role": "Co-founder and CEO", "bio": "Spent a decade at Sumo Logic and built its Warsaw engineering office. Also worked on NVIDIA CUDA and at Facebook.", "claim_ids": [ "quesma.claim.founders", "quesma.claim.audit-jacek-migda-background" ] }, { "person_id": "quesma.person.pawe-brzoska", "role": "Co-founder; advisor, former CPO", "bio": "Former Sumo Logic colleague. Left Quesma’s operational team in January 2025 and remained an advisor.", "claim_ids": [ "quesma.claim.founders", "quesma.claim.audit-pawe-brzoska-background" ] } ], "notes": [], "traction_heading": "Product and research activity", "metric_ids": [ "quesma.metric.oct14-0-customer-benchmark-result" ], "activity": { "title": "Benchmark audit", "description": "The 66.5% result belongs to Blitzy on SWE-bench Pro Public, using Blitzy’s testing setup. Quesma also publishes tests of AI software that writes code.", "claim_ids": [ "quesma.claim.oct14-metric-0-customer-benchmark-result", "quesma.claim.oct14-product-2" ] }, "metric_context": { "quesma.metric.oct14-0-customer-benchmark-result": { "scope": "Blitzy · SWE-bench Pro Public", "note": "Blitzy harness" } } }, { "company_id": "refresh", "products": [ { "name": "Computer-use software worlds", "description": "Realistic, connected applications for business tasks, including electronic health records.", "record_ids": [ "refresh.product.oct14-0" ], "claim_ids": [ "refresh.claim.oct14-product-0" ] }, { "name": "BlenderBench", "description": "Expert-built tasks for modeling and animation, including rigging and materials. Also covers physics simulation, with versioned checks of results.", "record_ids": [ "refresh.product.oct14-1" ], "claim_ids": [ "refresh.claim.oct14-product-1" ] }, { "name": "computer-1", "description": "A tool for testing computer use across models, contributed to Harbor.", "record_ids": [ "refresh.product.oct14-2" ], "claim_ids": [ "refresh.claim.oct14-product-2" ] }, { "name": "Trajectories", "description": "Step-by-step records of agent actions, with benchmark results.", "record_ids": [ "refresh.product.oct14-3" ], "claim_ids": [ "refresh.claim.oct14-product-3" ] } ], "relationships": [], "people": [ { "person_id": "refresh.person.christopher-settles", "role": "Co-founder and CEO", "bio": "", "claim_ids": [ "refresh.claim.founders" ] }, { "person_id": "refresh.person.erik-quintanilla", "role": "Co-founder and CTO", "bio": "", "claim_ids": [ "refresh.claim.founders" ] } ], "notes": [] }, { "company_id": "benchflow", "products": [ { "name": "SkillsBench", "description": "Tests whether reusable instructions improve agents’ task performance.", "record_ids": [ "benchflow.product.oct14-0" ], "claim_ids": [ "benchflow.claim.oct14-product-0" ] }, { "name": "ClawsBench", "description": "Replicas of Gmail and Calendar, plus Docs, Drive and Slack, that retain changes between actions. Scores capability and safety separately.", "record_ids": [ "benchflow.product.oct14-1" ], "claim_ids": [ "benchflow.claim.oct14-product-1" ] }, { "name": "Robo Use", "description": "Lets coding agents operate simulated robots through shell commands.", "record_ids": [ "benchflow.product.oct14-2" ], "claim_ids": [ "benchflow.claim.oct14-product-2" ] }, { "name": "PostTrain Arena", "description": "Community environments tested using a fixed training recipe. Training runs are currently paused.", "record_ids": [ "benchflow.product.oct14-3" ], "claim_ids": [ "benchflow.claim.oct14-product-3" ], "status": "Training paused" }, { "name": "BenchFlow runtime", "description": "Open-source tools for running and scoring agents in test environments.", "record_ids": [ "benchflow.product.oct14-4" ], "claim_ids": [ "benchflow.claim.oct14-product-4" ] } ], "relationships": [ { "name": "Hugging Face", "relationship": "compute_provider", "description": "Named computing provider for PostTrain. Paying-customer status is unverified.", "record_ids": [ "benchflow.customer.1" ], "claim_ids": [ "benchflow.claim.customers-1" ] }, { "name": "Fireworks", "relationship": "compute_provider", "description": "Named computing provider for PostTrain. Paying-customer status is unverified.", "record_ids": [ "benchflow.customer.1" ], "claim_ids": [ "benchflow.claim.customers-1" ] }, { "name": "Prime Intellect", "relationship": "compute_provider", "description": "Named computing provider for PostTrain. Paying-customer status is unverified.", "record_ids": [ "benchflow.customer.1" ], "claim_ids": [ "benchflow.claim.customers-1" ] }, { "name": "DeepMind", "relationship": "event_collaborator", "description": "Named event collaborator. A sales relationship is unverified.", "record_ids": [ "benchflow.customer.1" ], "claim_ids": [ "benchflow.claim.customers-1" ] } ], "people": [ { "person_id": "benchflow.person.xiangyi-li", "role": "Founder", "bio": "", "claim_ids": [ "benchflow.claim.founders" ] } ], "notes": [] }, { "company_id": "trajectory_labs_pbc", "products": [ { "name": "Safety evaluations and red teaming", "description": "Experts design new attacks to test weaknesses in advanced agents and their cybersecurity safeguards.", "record_ids": [ "trajectory_labs_pbc.product.oct14-0" ], "claim_ids": [ "trajectory_labs_pbc.claim.oct14-product-0" ] }, { "name": "Safety RL environments", "description": "Tasks and attack examples reveal failures and help train safer behavior.", "record_ids": [ "trajectory_labs_pbc.product.oct14-1" ], "claim_ids": [ "trajectory_labs_pbc.claim.oct14-product-1" ] }, { "name": "Search-PI Complex", "description": "A benchmark developed with Meta for prompt injection (malicious instructions that try to redirect an AI). Uses proprietary attacks generated by attacker models.", "record_ids": [ "trajectory_labs_pbc.product.oct14-2" ], "claim_ids": [ "trajectory_labs_pbc.claim.oct14-product-2" ] } ], "relationships": [ { "name": "Meta", "relationship": "research_collaborator", "description": "Meta confirms Search-PI Complex development and delivery of prompt-injection attacks. Payment is undisclosed.", "record_ids": [ "trajectory_labs_pbc.customer.0" ], "claim_ids": [ "trajectory_labs_pbc.claim.audit-meta-delivery" ] }, { "name": "Anthropic", "relationship": "commissioned_evaluation", "description": "Anthropic confirms external cyber testing and a commissioned prompt-injection evaluation. Contract value and payment are undisclosed.", "record_ids": [ "trajectory_labs_pbc.customer.1", "trajectory_labs_pbc.customer.audit-anthropic-commission" ], "claim_ids": [ "trajectory_labs_pbc.claim.customers-1", "trajectory_labs_pbc.claim.audit-anthropic-commission" ] } ], "people": [ { "person_id": "trajectory_labs_pbc.person.peter-mcintyre", "role": "Co-founder and CEO", "bio": "Former Centre for the Governance of AI research scholar and early 80,000 Hours employee.", "claim_ids": [ "trajectory_labs_pbc.claim.audit-peter-mcintyre-background", "trajectory_labs_pbc.claim.founders" ] }, { "person_id": "trajectory_labs_pbc.person.ben-west", "role": "Co-founder and CTO", "bio": "Former METR engineering head and author of time-horizons research.", "claim_ids": [ "trajectory_labs_pbc.claim.audit-ben-west-background", "trajectory_labs_pbc.claim.founders" ] }, { "person_id": "trajectory_labs_pbc.person.sami-jawhar", "role": "Founding team", "bio": "Former METR engineering head and research engineer with AI-safety publications.", "claim_ids": [ "trajectory_labs_pbc.claim.audit-sami-jawhar-background", "trajectory_labs_pbc.claim.founders" ] }, { "person_id": "trajectory_labs_pbc.person.ryan-carey", "role": "Founding team", "bio": "Former technical AI-safety lead at Oxford’s Future of Humanity Institute. Holds an Oxford statistics PhD.", "claim_ids": [ "trajectory_labs_pbc.claim.audit-ryan-carey-background", "trajectory_labs_pbc.claim.founders" ] } ], "notes": [] }, { "company_id": "elorian", "products": [ { "name": "Visual-thinking foundation model", "description": "An unreleased model for spatial and physical reasoning. Early access has a waitlist.", "record_ids": [ "elorian.product.oct14-0" ], "claim_ids": [ "elorian.claim.oct14-product-0" ], "status": "Unreleased; early-access waitlist" } ], "relationships": [], "people": [ { "person_id": "elorian.person.andrew-dai", "role": "Co-founder and CEO", "bio": "", "claim_ids": [ "elorian.claim.founders-0" ] }, { "person_id": "elorian.person.yinfei-yang", "role": "Co-founder and Chief Multimodal Architect", "bio": "", "claim_ids": [ "elorian.claim.founders-0" ] }, { "person_id": "elorian.person.seth-neel", "role": "Co-founder and Member of Technical Staff", "bio": "", "claim_ids": [ "elorian.claim.founders-0" ] } ], "notes": [] }, { "company_id": "a-team", "products": [ { "name": "Agentic Teammates", "description": "Teams build business AI agents that the client owns.", "record_ids": [ "a-team.product.oct14-0", "a-team.product.0", "a-team.product.1" ], "claim_ids": [ "a-team.claim.oct14-product-0", "a-team.claim.products-0", "a-team.claim.products-1" ] }, { "name": "Intelligence Layer", "description": "Senior engineering teams build and run business AI systems.", "record_ids": [ "a-team.product.oct14-1" ], "claim_ids": [ "a-team.claim.oct14-product-1" ] }, { "name": "Forward Deployed Engineers", "description": "Engineers build task evaluations and systems for further model training.", "record_ids": [ "a-team.product.oct14-2", "a-team.product.0", "a-team.product.1" ], "claim_ids": [ "a-team.claim.oct14-product-2", "a-team.claim.products-0", "a-team.claim.products-1" ] }, { "name": "Technical Talent", "description": "Technical specialists join hourly or ongoing work, including advisory engagements.", "record_ids": [ "a-team.product.oct14-3", "a-team.product.0", "a-team.product.1" ], "claim_ids": [ "a-team.claim.oct14-product-3", "a-team.claim.products-0", "a-team.claim.products-1" ] } ], "relationships": [ { "name": "Blank Street", "relationship": "claimed_customer", "description": "A.Team’s case study describes an ordering app in live use and three years of product-engineering work.", "record_ids": [ "a-team.customer.0", "a-team.customer.2" ], "claim_ids": [ "a-team.claim.customers-0", "a-team.claim.customers-2" ] }, { "name": "Apprentice", "relationship": "claimed_customer", "description": "Named in A.Team’s case studies. Revenue from the work is not independently verified.", "record_ids": [ "a-team.customer.0" ], "claim_ids": [ "a-team.claim.customers-0" ] }, { "name": "Chief", "relationship": "claimed_customer", "description": "Named in A.Team’s case studies. Revenue from the work is not independently verified.", "record_ids": [ "a-team.customer.0" ], "claim_ids": [ "a-team.claim.customers-0" ] }, { "name": "D-ID", "relationship": "claimed_customer", "description": "Named in A.Team’s case studies. Revenue from the work is not independently verified.", "record_ids": [ "a-team.customer.0" ], "claim_ids": [ "a-team.claim.customers-0" ] }, { "name": "EY", "relationship": "claimed_customer", "description": "Named in A.Team’s case studies. Revenue from the work is not independently verified.", "record_ids": [ "a-team.customer.0" ], "claim_ids": [ "a-team.claim.customers-0" ] }, { "name": "Accrue", "relationship": "claimed_customer", "description": "Named in A.Team’s case studies. Revenue from the work is not independently verified.", "record_ids": [ "a-team.customer.0" ], "claim_ids": [ "a-team.claim.customers-0" ] } ], "people": [ { "person_id": "a-team.person.raphael-ouzan", "role": "Co-founder", "bio": "", "claim_ids": [ "a-team.claim.founders-0" ] }, { "person_id": "a-team.person.kobi-mastri", "role": "Co-founder", "bio": "", "claim_ids": [ "a-team.claim.founders-0" ] } ], "notes": [] }, { "company_id": "aixi", "products": [ { "name": "AIXI safety research", "description": "Connects mathematical theories of AI agents with practical research on model safety.", "record_ids": [ "aixi.product.0", "aixi.product.oct14-0" ], "claim_ids": [ "aixi.claim.products-0", "aixi.claim.oct14-product-0" ] }, { "name": "Research Fellowship", "description": "About three months of research, with supervision and computing support.", "record_ids": [ "aixi.product.1", "aixi.product.oct14-1" ], "claim_ids": [ "aixi.claim.products-1", "aixi.claim.oct14-product-1" ] } ], "relationships": [ { "name": "Principles of Intelligence", "relationship": "fiscal_sponsor", "description": "The Minnesota nonprofit serves as fiscal sponsor (a nonprofit that manages funds for another project).", "record_ids": [], "claim_ids": [ "aixi.claim.audit-aixi-entity" ] } ], "people": [ { "person_id": "aixi.person.cole-wyeth", "role": "Founder and Executive Director", "bio": "Pursued computer-science PhD research at Waterloo under Ming Li, with advice from Marcus Hutter. Holds an MS in mathematics from Minnesota.", "claim_ids": [ "aixi.claim.founders", "aixi.claim.audit-aixi-founder-background" ] } ], "notes": [ { "text": "Marcus Hutter serves as part-time Research Director. Aram Ebtekar is Founding Research Scientist.", "claim_ids": [ "aixi.claim.founders", "aixi.claim.audit-aixi-entity" ] } ], "traction_heading": "Research activity", "metric_ids": [], "activity": { "title": "Foundations of agent safety", "description": "Research connects mathematical ideas about programs and information with tests of AI model safety. The fellowship provides supervision and computing support.", "claim_ids": [ "aixi.claim.products-0", "aixi.claim.products-1" ] } }, { "company_id": "alinia", "products": [ { "name": "Alinia Judges", "description": "Models trained to judge outputs against institutional and legal policies.", "record_ids": [ "alinia.product.oct14-0", "alinia.product.0" ], "claim_ids": [ "alinia.claim.oct14-product-0", "alinia.claim.products-0" ] }, { "name": "Guardrails API", "description": "Applies policy checks while AI systems are running.", "record_ids": [ "alinia.product.oct14-1", "alinia.product.0" ], "claim_ids": [ "alinia.claim.oct14-product-1", "alinia.claim.products-0" ] }, { "name": "AI Compliance Platform", "description": "Expert review and audit records support deployment into live use.", "record_ids": [ "alinia.product.oct14-2", "alinia.product.0" ], "claim_ids": [ "alinia.claim.oct14-product-2", "alinia.claim.products-0" ] }, { "name": "Seny", "description": "A specialized legal-compliance model released in October 2026.", "record_ids": [ "alinia.product.oct14-3", "alinia.product.1" ], "claim_ids": [ "alinia.claim.oct14-product-3", "alinia.claim.products-1" ] } ], "relationships": [ { "name": "Santander Group", "relationship": "claimed_customer", "description": "Named in the December 2025 seed announcement, with named executive comments. Contract fees and deployment scope are undisclosed.", "record_ids": [ "alinia.customer.0" ], "claim_ids": [ "alinia.claim.customers-0" ] }, { "name": "Origin Financial", "relationship": "claimed_customer", "description": "Named in the December 2025 seed announcement, with named executive comments. Contract fees and deployment scope are undisclosed.", "record_ids": [ "alinia.customer.0" ], "claim_ids": [ "alinia.claim.customers-0" ] }, { "name": "Barcelona Supercomputing Center", "relationship": "compute_support", "description": "Alinia reports research and computing support for Seny through EuroHPC AI Factory. Purchasing is unverified.", "record_ids": [ "alinia.customer.1" ], "claim_ids": [ "alinia.claim.customers-1", "alinia.claim.audit-alinia-bsc-tag" ] } ], "people": [ { "person_id": "alinia.person.ariadna-font-llitj-s", "role": "Co-founder and CEO", "bio": "", "claim_ids": [ "alinia.claim.founders" ] }, { "person_id": "alinia.person.carlos-mu-oz-ferrandis", "role": "Co-founder and COO", "bio": "", "claim_ids": [ "alinia.claim.founders" ] } ], "notes": [] }, { "company_id": "apto", "products": [ { "name": "harBest Annotation", "description": "Managed data collection and labeling.", "record_ids": [ "apto.product.oct14-0", "apto.product.0" ], "claim_ids": [ "apto.claim.oct14-product-0", "apto.claim.products-0" ] }, { "name": "harBest Expert", "description": "Specialists create instruction data and evaluate model quality.", "record_ids": [ "apto.product.oct14-1", "apto.product.0" ], "claim_ids": [ "apto.claim.oct14-product-1", "apto.claim.products-0" ] }, { "name": "AI Solutions", "description": null, "record_ids": [ "apto.product.oct14-2" ], "claim_ids": [ "apto.claim.oct14-product-2" ] }, { "name": "AI Datasets", "description": null, "record_ids": [ "apto.product.oct14-3" ], "claim_ids": [ "apto.claim.oct14-product-3" ] } ], "relationships": [ { "name": "Micro Control Systems", "relationship": "claimed_customer", "description": "A customer interview on Aptō’s site describes paid factory-image labeling and use in production. It reports improved image recognition.", "record_ids": [ "apto.customer.0", "apto.customer.audit-apto-case-outcome" ], "claim_ids": [ "apto.claim.customers-0", "apto.claim.audit-apto-case-classification-apto.customer.0", "apto.claim.audit-apto-case-outcome" ] }, { "name": "RIKEN", "relationship": "reported_customer", "description": "An investor interview reports RIKEN use. Contract scope and payment are unverified.", "record_ids": [ "apto.customer.1" ], "claim_ids": [ "apto.claim.customers-1" ] }, { "name": "ORIX", "relationship": "reported_customer", "description": "An investor interview reports ORIX use. Contract scope and payment are unverified.", "record_ids": [ "apto.customer.1" ], "claim_ids": [ "apto.claim.customers-1" ] }, { "name": "Cohere", "relationship": "displayed_logo", "description": "Aptō displays Cohere’s logo. Paid contract scope is unverified.", "record_ids": [ "apto.customer.1" ], "claim_ids": [ "apto.claim.customers-1" ] }, { "name": "Fujitsu", "relationship": "displayed_logo", "description": "Aptō displays Fujitsu’s logo. Paid contract scope is unverified.", "record_ids": [ "apto.customer.1" ], "claim_ids": [ "apto.claim.customers-1" ] }, { "name": "SoftBank", "relationship": "displayed_logo", "description": "Aptō displays SoftBank’s logo. Paid contract scope is unverified.", "record_ids": [ "apto.customer.1" ], "claim_ids": [ "apto.claim.customers-1" ] }, { "name": "Canon", "relationship": "displayed_logo", "description": "Aptō displays Canon’s logo. Paid contract scope is unverified.", "record_ids": [ "apto.customer.1" ], "claim_ids": [ "apto.claim.customers-1" ] }, { "name": "NTT Data", "relationship": "displayed_logo", "description": "Aptō displays NTT Data’s logo. Paid contract scope is unverified.", "record_ids": [ "apto.customer.1" ], "claim_ids": [ "apto.claim.customers-1" ] } ], "people": [ { "person_id": "apto.person.ryo-takashina", "role": "Founder and CEO", "bio": "Worked in backend development and IT consulting. Founded APTO in 2020.", "claim_ids": [ "apto.claim.founders", "apto.claim.audit-apto-founder-background" ] } ], "notes": [] }, { "company_id": "g2i", "products": [ { "name": "RLHF Pipelines", "description": "Engineers build human-feedback systems through teams embedded with clients or managed projects.", "record_ids": [ "g2i.product.oct14-0", "g2i.product.0" ], "claim_ids": [ "g2i.claim.oct14-product-0", "g2i.claim.products-0" ] }, { "name": "SFT Datasets", "description": "Supervised fine-tuning datasets (examples that teach models to produce desired responses).", "record_ids": [ "g2i.product.oct14-1" ], "claim_ids": [ "g2i.claim.oct14-product-1" ] }, { "name": "Evaluations", "description": "Tests of model performance.", "record_ids": [ "g2i.product.oct14-2" ], "claim_ids": [ "g2i.claim.oct14-product-2" ] }, { "name": "RL Environments", "description": "Custom training environments with evaluations and grading rules.", "record_ids": [ "g2i.product.oct14-3", "g2i.product.0" ], "claim_ids": [ "g2i.claim.oct14-product-3", "g2i.claim.products-0" ] }, { "name": "Seed Data", "description": null, "record_ids": [ "g2i.product.oct14-4" ], "claim_ids": [ "g2i.claim.oct14-product-4" ] } ], "relationships": [ { "name": "Automattic", "relationship": "claimed_customer", "description": "G2i names Automattic as a staffing reference. AI-data purchasing is unverified.", "record_ids": [ "g2i.customer.0" ], "claim_ids": [ "g2i.claim.customers-0" ] }, { "name": "Discover Card", "relationship": "claimed_customer", "description": "G2i names Discover Card as a staffing reference. AI-data purchasing is unverified.", "record_ids": [ "g2i.customer.0" ], "claim_ids": [ "g2i.claim.customers-0" ] }, { "name": "Webflow", "relationship": "claimed_customer", "description": "G2i names Webflow as a staffing reference. AI-data purchasing is unverified.", "record_ids": [ "g2i.customer.0" ], "claim_ids": [ "g2i.claim.customers-0" ] }, { "name": "Shopmonkey", "relationship": "claimed_customer", "description": "G2i’s case study includes engineering-leader comments describing 20 hires and staffing services.", "record_ids": [ "g2i.customer.2" ], "claim_ids": [ "g2i.claim.customers-2" ] } ], "people": [ { "person_id": "g2i.person.gabe-greenberg", "role": "Founder and CEO", "bio": "", "claim_ids": [ "g2i.claim.founders-0" ] } ], "notes": [] }, { "company_id": "giskard", "products": [ { "name": "Continuous Red Teaming", "description": "Deliberate attacks test security risks and failures specific to a business.", "record_ids": [ "giskard.product.oct14-0", "giskard.product.0" ], "claim_ids": [ "giskard.claim.oct14-product-0", "giskard.claim.products-0" ] }, { "name": "LLM Evaluation", "description": "Automated checks and expert review assess model behavior.", "record_ids": [ "giskard.product.oct14-1", "giskard.product.0" ], "claim_ids": [ "giskard.claim.oct14-product-1", "giskard.claim.products-0" ] }, { "name": "AI Guardrails", "description": null, "record_ids": [ "giskard.product.oct14-2" ], "claim_ids": [ "giskard.claim.oct14-product-2" ] }, { "name": "Giskard Hub", "description": "A platform for businesses to test and review AI.", "record_ids": [ "giskard.product.oct14-3", "giskard.product.0" ], "claim_ids": [ "giskard.claim.oct14-product-3", "giskard.claim.products-0" ] }, { "name": "Phare LLM Benchmark", "description": null, "record_ids": [ "giskard.product.oct14-4" ], "claim_ids": [ "giskard.claim.oct14-product-4" ] } ], "relationships": [ { "name": "Michelin", "relationship": "claimed_customer", "description": "Named comments on Giskard’s site describe its use in Michelin’s language-model evaluations.", "record_ids": [ "giskard.customer.0" ], "claim_ids": [ "giskard.claim.customers-0" ] }, { "name": "BNP Paribas", "relationship": "claimed_customer", "description": "Giskard hosts attributable customer testimony from BNP Paribas. Contract value is undisclosed.", "record_ids": [ "giskard.customer.0" ], "claim_ids": [ "giskard.claim.customers-0" ] }, { "name": "Decathlon", "relationship": "claimed_customer", "description": "Giskard hosts attributable customer testimony from Decathlon. Contract value is undisclosed.", "record_ids": [ "giskard.customer.0" ], "claim_ids": [ "giskard.claim.customers-0" ] }, { "name": "Google DeepMind", "relationship": "research_collaborator", "description": "Giskard names a Phare research collaboration. Paying software-customer status is unverified.", "record_ids": [ "giskard.customer.1" ], "claim_ids": [ "giskard.claim.customers-1", "giskard.claim.audit-giskard-partners-not-customers" ] }, { "name": "European Commission", "relationship": "funding_support", "description": "Giskard names funding support for Phare research. Paying software-customer status is unverified.", "record_ids": [ "giskard.customer.1" ], "claim_ids": [ "giskard.claim.customers-1", "giskard.claim.audit-giskard-partners-not-customers" ] }, { "name": "Bpifrance", "relationship": "funding_support", "description": "Giskard names funding support for Phare research. Paying software-customer status is unverified.", "record_ids": [ "giskard.customer.1" ], "claim_ids": [ "giskard.claim.customers-1", "giskard.claim.audit-giskard-partners-not-customers" ] } ], "people": [ { "person_id": "giskard.person.alex-combessie", "role": "Co-founder and co-CEO", "bio": "", "claim_ids": [ "giskard.claim.founders" ] }, { "person_id": "giskard.person.jean-marie-john-mathews", "role": "Co-founder and co-CEO", "bio": "", "claim_ids": [ "giskard.claim.founders" ] }, { "person_id": "giskard.person.andrey-avtomonov", "role": "Co-founder", "bio": "", "claim_ids": [ "giskard.claim.founders" ] } ], "notes": [] }, { "company_id": "hackerrank", "products": [ { "name": "HackerRank Bench", "description": "Private, extended engineering tasks. Hand-built checks score how the agent behaves.", "record_ids": [ "hackerrank.product.oct14-0", "hackerrank.product.1" ], "claim_ids": [ "hackerrank.claim.oct14-product-0", "hackerrank.claim.products-1" ] }, { "name": "ASTRA benchmark", "description": "Extended engineering tasks in environments that can be reproduced.", "record_ids": [ "hackerrank.product.oct14-1", "hackerrank.product.0" ], "claim_ids": [ "hackerrank.claim.oct14-product-1", "hackerrank.claim.products-0" ] }, { "name": "Coding datasets", "description": "Coding examples for training models.", "record_ids": [ "hackerrank.product.oct14-2" ], "claim_ids": [ "hackerrank.claim.oct14-product-2" ] }, { "name": "Private benchmarking", "description": "Private tests of model performance.", "record_ids": [ "hackerrank.product.oct14-3" ], "claim_ids": [ "hackerrank.claim.oct14-product-3" ] }, { "name": "RL environments", "description": "Reinforcement learning environments (models learn by trying tasks and receiving scores).", "record_ids": [ "hackerrank.product.oct14-4" ], "claim_ids": [ "hackerrank.claim.oct14-product-4" ] } ], "relationships": [ { "name": "Airbnb", "relationship": "historical_customer", "description": "Named as a past customer of HackerRank. Astra customer status is unverified.", "record_ids": [ "hackerrank.customer.1" ], "claim_ids": [ "hackerrank.claim.customers-1" ] }, { "name": "Box", "relationship": "historical_customer", "description": "Named as a past customer of HackerRank. Astra customer status is unverified.", "record_ids": [ "hackerrank.customer.1" ], "claim_ids": [ "hackerrank.claim.customers-1" ] }, { "name": "Amazon", "relationship": "historical_customer", "description": "Named as a past customer of HackerRank. Astra customer status is unverified.", "record_ids": [ "hackerrank.customer.1" ], "claim_ids": [ "hackerrank.claim.customers-1" ] } ], "people": [ { "person_id": "hackerrank.person.vivek-ravisankar", "role": "HackerRank co-founder and CEO", "bio": "", "claim_ids": [ "hackerrank.claim.founders-0" ] }, { "person_id": "hackerrank.person.harishankaran-karunanidhi", "role": "HackerRank co-founder and CTO", "bio": "", "claim_ids": [ "hackerrank.claim.founders-0" ] } ], "notes": [] }, { "company_id": "laude", "products": [ { "name": "Terminal-Bench", "description": "Open tests of agent work in command-line environments (text-based interfaces for running commands on computers). The reviewed homepage shows version 4.0.", "record_ids": [ "laude.product.0", "laude.product.oct14-0" ], "claim_ids": [ "laude.claim.products-0", "laude.claim.oct14-product-0" ] }, { "name": "Slingshots", "description": "Grants and hands-on support for research projects.", "record_ids": [ "laude.product.0", "laude.product.oct14-2" ], "claim_ids": [ "laude.claim.products-0", "laude.claim.oct14-product-2" ] }, { "name": "Moonshots", "description": "Support for research labs pursuing work over longer periods.", "record_ids": [ "laude.product.0", "laude.product.oct14-3" ], "claim_ids": [ "laude.claim.products-0", "laude.claim.oct14-product-3" ] }, { "name": "Harbor", "description": null, "record_ids": [ "laude.product.oct14-1" ], "claim_ids": [ "laude.claim.oct14-product-1" ] } ], "relationships": [ { "name": "Anthropic", "relationship": "benchmark_adoption", "description": "Anthropic reports using Terminal-Bench to evaluate Claude.", "record_ids": [ "laude.customer.0" ], "claim_ids": [ "laude.claim.customers-0", "laude.claim.audit-laude-noncustomers-laude.customer.0" ] } ], "people": [ { "person_id": "laude.person.andy-konwinski", "role": "Co-founder", "bio": "His UC Berkeley PhD work on Hadoop led into Apache Spark and Databricks. He also helped build Perplexity.", "claim_ids": [ "laude.claim.founders-0", "laude.claim.audit-laude-founder-background" ] }, { "person_id": "laude.person.chris-rytting", "role": "Co-founder", "bio": "", "claim_ids": [ "laude.claim.founders-0" ] }, { "person_id": "laude.person.k.-tighe", "role": "Co-founder", "bio": "", "claim_ids": [ "laude.claim.founders-0" ] }, { "person_id": "laude.person.justin-fiedler", "role": "Co-founder", "bio": "", "claim_ids": [ "laude.claim.founders-0" ] }, { "person_id": "laude.person.lindsey-gregory", "role": "Co-founder", "bio": "", "claim_ids": [ "laude.claim.founders-0" ] } ], "notes": [ { "text": "Andy Konwinski pledged $100 million in donations. How much cash the Institute received is unverified.", "claim_ids": [ "laude.claim.funding_rounds-0", "laude.claim.total_funding", "laude.claim.audit-laude-pledge" ] }, { "text": "The Institute reports $4.1 million in seed grants (early funding to help research projects get started) across 25 teams.", "claim_ids": [ "laude.claim.audit-laude-deployed-grants-metric" ] } ], "traction_heading": "Research and grantmaking", "metric_ids": [ "laude.metric.audit-laude-deployed-grants", "laude.metric.oct14-0-benchmark-dataset-size" ], "activity": { "title": "Open research support", "description": "Slingshots and Moonshots support research teams. Terminal-Bench tests how well AI software completes computer tasks using text commands.", "claim_ids": [ "laude.claim.products-0" ] }, "metric_context": { "laude.metric.audit-laude-deployed-grants": { "scope": "Grants paid to research teams" }, "laude.metric.oct14-0-benchmark-dataset-size": { "scope": "Terminal-Bench 2.0 · historical" } } }, { "company_id": "prime_intellect", "products": [ { "name": "Lab", "description": "A platform for reinforcement learning (models learn by trying tasks and receiving scores) and further model training.", "record_ids": [ "prime_intellect.product.oct14-0", "prime_intellect.product.0" ], "claim_ids": [ "prime_intellect.claim.oct14-product-0", "prime_intellect.claim.products" ] }, { "name": "Environment Hub", "description": "Environments for model training and evaluation.", "record_ids": [ "prime_intellect.product.oct14-1", "prime_intellect.product.0" ], "claim_ids": [ "prime_intellect.claim.oct14-product-1", "prime_intellect.claim.products" ] }, { "name": "Verifiers", "description": null, "record_ids": [ "prime_intellect.product.oct14-2" ], "claim_ids": [ "prime_intellect.claim.oct14-product-2" ] }, { "name": "Prime-RL", "description": null, "record_ids": [ "prime_intellect.product.oct14-3" ], "claim_ids": [ "prime_intellect.claim.oct14-product-3" ] }, { "name": "Prime Inference Gateway", "description": null, "record_ids": [ "prime_intellect.product.oct14-4" ], "claim_ids": [ "prime_intellect.claim.oct14-product-4" ] }, { "name": "Sandboxes", "description": "Isolated environments for training and evaluation.", "record_ids": [ "prime_intellect.product.oct14-5", "prime_intellect.product.0" ], "claim_ids": [ "prime_intellect.claim.oct14-product-5", "prime_intellect.claim.products" ] }, { "name": "INTELLECT", "description": "Open model research from Prime Intellect.", "record_ids": [ "prime_intellect.product.oct14-6", "prime_intellect.product.0" ], "claim_ids": [ "prime_intellect.claim.oct14-product-6", "prime_intellect.claim.products" ] } ], "relationships": [ { "name": "Ramp", "relationship": "claimed_customer", "description": "Prime Intellect’s case study describes training the search component of Ramp’s FastAsk agent. Independent buyer confirmation and paid terms are unverified.", "record_ids": [ "prime_intellect.customer.0", "prime_intellect.customer.audit-pi-ramp-proof" ], "claim_ids": [ "prime_intellect.claim.audit-pi-ramp-classification-prime_intellect.customer.0", "prime_intellect.claim.audit-pi-ramp-proof" ] }, { "name": "Langdock", "relationship": "product_endorsement", "description": "Named product testimonial. Paid terms are unconfirmed.", "record_ids": [ "prime_intellect.customer.1" ], "claim_ids": [ "prime_intellect.claim.customers-1" ] } ], "people": [ { "person_id": "prime_intellect.person.vincent-weisser", "role": "Co-founder", "bio": "", "claim_ids": [ "prime_intellect.claim.founders" ] }, { "person_id": "prime_intellect.person.johannes-hagemann", "role": "Co-founder", "bio": "", "claim_ids": [ "prime_intellect.claim.founders" ] } ], "notes": [] }, { "company_id": "prolific", "products": [ { "name": "Prolific platform", "description": "Recruits verified participants for research studies.", "record_ids": [ "prolific.product.oct14-0", "prolific.product.0" ], "claim_ids": [ "prolific.claim.oct14-product-0", "prolific.claim.products-0" ] }, { "name": "Human Feedback API", "description": "Connects software directly to human feedback for AI training and evaluation.", "record_ids": [ "prolific.product.oct14-1", "prolific.product.0" ], "claim_ids": [ "prolific.claim.oct14-product-1", "prolific.claim.products-0" ] }, { "name": "HUMAINE", "description": null, "record_ids": [ "prolific.product.oct14-2" ], "claim_ids": [ "prolific.claim.oct14-product-2" ] }, { "name": "Prolific Expert Network", "description": "Groups of experts recruited for specialist research.", "record_ids": [ "prolific.product.oct14-3", "prolific.product.0" ], "claim_ids": [ "prolific.claim.oct14-product-3", "prolific.claim.products-0" ] }, { "name": "Managed Research Services", "description": "Managed projects that collect human feedback and evaluate models.", "record_ids": [ "prolific.product.oct14-4", "prolific.product.0" ], "claim_ids": [ "prolific.claim.oct14-product-4", "prolific.claim.products-0" ] } ], "relationships": [ { "name": "Ai2", "relationship": "claimed_customer", "description": "Prolific’s case study includes named Ai2 comments about delivered work on data across multiple formats.", "record_ids": [ "prolific.customer.0" ], "claim_ids": [ "prolific.claim.customers-0" ] }, { "name": "Google", "relationship": "historical_customer", "description": "Prolific’s 2023 financing release names Google as a customer/user. Current relationship status is unconfirmed.", "record_ids": [ "prolific.customer.1" ], "claim_ids": [ "prolific.claim.customers-1" ] }, { "name": "Stanford", "relationship": "historical_customer", "description": "Prolific’s 2023 financing release names Stanford as a customer/user. Current relationship status is unconfirmed.", "record_ids": [ "prolific.customer.1" ], "claim_ids": [ "prolific.claim.customers-1" ] }, { "name": "Oxford", "relationship": "historical_customer", "description": "Prolific’s 2023 financing release names Oxford as a customer/user. Current relationship status is unconfirmed.", "record_ids": [ "prolific.customer.1" ], "claim_ids": [ "prolific.claim.customers-1" ] }, { "name": "King’s College London", "relationship": "historical_customer", "description": "Prolific’s 2023 financing release names King’s College London as a customer/user. Current relationship status is unconfirmed.", "record_ids": [ "prolific.customer.1" ], "claim_ids": [ "prolific.claim.customers-1" ] }, { "name": "European Commission", "relationship": "historical_customer", "description": "Prolific’s 2023 financing release names European Commission as a customer/user. Current relationship status is unconfirmed.", "record_ids": [ "prolific.customer.1" ], "claim_ids": [ "prolific.claim.customers-1" ] } ], "people": [ { "person_id": "prolific.person.phelim-bradley", "role": "Co-founder", "bio": "", "claim_ids": [ "prolific.claim.founders" ] }, { "person_id": "prolific.person.ekaterina-damer", "role": "Co-founder", "bio": "", "claim_ids": [ "prolific.claim.founders" ] } ], "notes": [] }, { "company_id": "rws_trainai", "products": [ { "name": "TrainAI", "description": "Managed training data in multiple languages, with expert feedback.", "record_ids": [ "rws_trainai.product.oct14-0", "rws_trainai.product.0" ], "claim_ids": [ "rws_trainai.claim.oct14-product-0", "rws_trainai.claim.products-0" ] }, { "name": "TrainAI Community", "description": "A screened specialist network for AI data work.", "record_ids": [ "rws_trainai.product.oct14-1", "rws_trainai.product.0" ], "claim_ids": [ "rws_trainai.claim.oct14-product-1", "rws_trainai.claim.products-0" ] }, { "name": "AI data consulting", "description": null, "record_ids": [ "rws_trainai.product.oct14-2" ], "claim_ids": [ "rws_trainai.claim.oct14-product-2" ] }, { "name": "Generative AI training and fine-tuning", "description": "Expert data for model training, including human feedback and carefully designed prompts.", "record_ids": [ "rws_trainai.product.oct14-3", "rws_trainai.product.0" ], "claim_ids": [ "rws_trainai.claim.oct14-product-3", "rws_trainai.claim.products-0" ] } ], "relationships": [ { "name": "Cohere", "relationship": "strategic_partner", "description": "Strategic partner for RWS Language Weaver Pro. TrainAI customer status is unverified.", "record_ids": [ "rws_trainai.customer.1" ], "claim_ids": [ "rws_trainai.claim.customers-1" ] } ], "people": [ { "person_id": "rws_trainai.person.vasagi-kothandapani", "role": "TrainAI CEO", "bio": "Previously led an AI-data-services business at Appen. Held senior roles at Cognizant and CoreLogic.", "claim_ids": [ "rws_trainai.claim.audit-rws-leadership", "rws_trainai.claim.audit-rws-leadership-background" ] } ], "notes": [] }, { "company_id": "terac", "products": [ { "name": "Terac Expert Network", "description": "Recruits and verifies experts for AI training programs.", "record_ids": [ "terac.product.oct14-0", "terac.product.0" ], "claim_ids": [ "terac.claim.oct14-product-0", "terac.claim.products-0" ] }, { "name": "Terac API", "description": "Connects software directly to expert labor services.", "record_ids": [ "terac.product.oct14-1", "terac.product.0" ], "claim_ids": [ "terac.claim.oct14-product-1", "terac.claim.products-0" ] }, { "name": "Agent MCP", "description": null, "record_ids": [ "terac.product.oct14-2" ], "claim_ids": [ "terac.claim.oct14-product-2" ] }, { "name": "AI interviewer", "description": "AI screening and interviews for expert recruitment.", "record_ids": [ "terac.product.oct14-3", "terac.product.0" ], "claim_ids": [ "terac.claim.oct14-product-3", "terac.claim.products-0" ] }, { "name": "Terac ChatGPT App", "description": null, "record_ids": [ "terac.product.oct14-4" ], "claim_ids": [ "terac.claim.oct14-product-4" ] } ], "relationships": [ { "name": "Dialogue", "relationship": "claimed_customer", "description": "Terac’s case study describes recruiting 25 startup founders in under four days. Contract value is undisclosed.", "record_ids": [ "terac.customer.0" ], "claim_ids": [ "terac.claim.customers-0" ] }, { "name": "Maze", "relationship": "claimed_customer", "description": "Terac hosts an attributed testimonial from Maze. Payment is unverified.", "record_ids": [ "terac.customer.1" ], "claim_ids": [ "terac.claim.customers-1" ] }, { "name": "Spade", "relationship": "claimed_customer", "description": "Terac hosts an attributed testimonial from Spade. Payment is unverified.", "record_ids": [ "terac.customer.1" ], "claim_ids": [ "terac.claim.customers-1" ] }, { "name": "Vetto", "relationship": "claimed_customer", "description": "Terac hosts an attributed testimonial from Vetto. Payment is unverified.", "record_ids": [ "terac.customer.1" ], "claim_ids": [ "terac.claim.customers-1" ] }, { "name": "Patronus AI", "relationship": "displayed_logo", "description": "The logo appears on Terac’s homepage. A paid relationship is unverified.", "record_ids": [ "terac.customer.1" ], "claim_ids": [ "terac.claim.customers-1" ] } ], "people": [ { "person_id": "terac.person.zac-baker", "role": "Co-founder and CEO", "bio": "", "claim_ids": [ "terac.claim.founders-0" ] }, { "person_id": "terac.person.jack-blair", "role": "Co-founder and CTO", "bio": "", "claim_ids": [ "terac.claim.founders-0" ] } ], "notes": [] }, { "company_id": "toloka", "products": [ { "name": "Toloka Arena", "description": null, "record_ids": [ "toloka.product.oct14-0" ], "claim_ids": [ "toloka.claim.oct14-product-0" ] }, { "name": "Agent Data", "description": "Expert task records and human evaluation for agent training.", "record_ids": [ "toloka.product.oct14-1", "toloka.product.major-0" ], "claim_ids": [ "toloka.claim.oct14-product-1", "toloka.claim.major-products-0" ] }, { "name": "Coding Data", "description": "Coding datasets written by experts for model training.", "record_ids": [ "toloka.product.oct14-2", "toloka.product.major-0" ], "claim_ids": [ "toloka.claim.oct14-product-2", "toloka.claim.major-products-0" ] }, { "name": "Physical AI", "description": null, "record_ids": [ "toloka.product.oct14-3" ], "claim_ids": [ "toloka.claim.oct14-product-3" ] }, { "name": "Mindrift", "description": null, "record_ids": [ "toloka.product.oct14-4" ], "claim_ids": [ "toloka.claim.oct14-product-4" ] }, { "name": "Tendem", "description": null, "record_ids": [ "toloka.product.oct14-5" ], "claim_ids": [ "toloka.claim.oct14-product-5" ] } ], "relationships": [ { "name": "Shopify", "relationship": "customer", "description": "Shopify engineering confirms that Toloka experts correct and score conversations for Sidekick’s ongoing training.", "record_ids": [ "toloka.customer.major-0" ], "claim_ids": [ "toloka.claim.major-customers-0" ] }, { "name": "Anthropic", "relationship": "claimed_customer", "description": "Named as a customer in Toloka’s investment announcement. Separate buyer confirmation was not reviewed.", "record_ids": [ "toloka.customer.major-1" ], "claim_ids": [ "toloka.claim.major-customers-1" ] }, { "name": "Amazon", "relationship": "claimed_customer", "description": "Named as a customer in Toloka’s investment announcement. Separate buyer confirmation was not reviewed.", "record_ids": [ "toloka.customer.major-1" ], "claim_ids": [ "toloka.claim.major-customers-1" ] }, { "name": "Microsoft", "relationship": "claimed_customer", "description": "Named as a customer in Toloka’s investment announcement. Separate buyer confirmation was not reviewed.", "record_ids": [ "toloka.customer.major-1" ], "claim_ids": [ "toloka.claim.major-customers-1" ] }, { "name": "poolside", "relationship": "claimed_customer", "description": "Named as a customer in Toloka’s investment announcement. Separate buyer confirmation was not reviewed.", "record_ids": [ "toloka.customer.major-1" ], "claim_ids": [ "toloka.claim.major-customers-1" ] }, { "name": "Recraft", "relationship": "claimed_customer", "description": "Named as a customer in Toloka’s investment announcement. Separate buyer confirmation was not reviewed.", "record_ids": [ "toloka.customer.major-1" ], "claim_ids": [ "toloka.claim.major-customers-1" ] } ], "people": [ { "person_id": "toloka.person.olga-megorskaya", "role": "Founder", "bio": "", "claim_ids": [ "toloka.claim.major-founders-0" ] } ], "notes": [] }, { "company_id": "afterquery", "summary": "Experts create data and practice environments to train AI models. AfterQuery also builds custom models and AI tools that carry out tasks for businesses.", "product_heading": "Training data products", "featured_case": { "title": "Inside NVIDIA’s office-task training", "eyebrow": "Dataset use", "description": "NVIDIA used AfterQuery tasks to train models for office work. Model-generated action records supported initial student-model training and later training with rewards.", "qualifier": "NVIDIA confirms task use. Payment and contract value are undisclosed.", "claim_ids": [ "afterquery.claim.round2-relationship-0" ] }, "products": [ { "name": "Office Agent Training Dataset", "description": "Professional assignments based on files. Tasks require analysis across several steps, with graded outputs.", "record_ids": [ "afterquery.product.round2-office-agent-data", "afterquery.product.oct14-0", "afterquery.product.oct14-1" ], "claim_ids": [ "afterquery.claim.round2-product-office-agent-data", "afterquery.claim.oct14-product-0", "afterquery.claim.oct14-product-1" ], "aliases": [ "Off-the-shelf Data", "Off-The-Shelf Office Agent Training Dataset" ] }, { "name": "SWE agent training datasets", "description": "Experts write tasks from private business codebases. Automated tests check the results.", "record_ids": [ "afterquery.product.round2-swe-agent-data" ], "claim_ids": [ "afterquery.claim.round2-product-swe-agent-data" ] }, { "name": "Terminal-agent training data", "description": "Successful action records from command-line tasks (text commands used to operate a computer). Tasks run in isolated environments and earn rewards from tests.", "record_ids": [ "afterquery.product.round2-terminal-agent-data" ], "claim_ids": [ "afterquery.claim.round2-product-terminal-agent-data" ] }, { "name": "Customer-service tool-use data", "description": "Service tasks with several steps. Checks confirm whether calls to software tools work correctly.", "record_ids": [ "afterquery.product.round2-customer-service-tools" ], "claim_ids": [ "afterquery.claim.round2-product-customer-service-tools" ] }, { "name": "Custom RL environments", "description": "Training tasks using connected software tools or browser and desktop applications. Expert action records and grading support reinforcement learning (models learn by trying tasks and receiving scores).", "record_ids": [ "afterquery.product.round2-custom-environments", "afterquery.product.oct14-2", "afterquery.product.oct14-3" ], "claim_ids": [ "afterquery.claim.round2-product-custom-environments", "afterquery.claim.oct14-product-2", "afterquery.claim.oct14-product-3" ], "aliases": [ "Tool-calling RL Environments", "Computer-use and Browser-use Environments" ] }, { "name": "Expert demonstrations and feedback", "description": "Examples teach desired responses, with comparisons showing which answers experts prefer. Expert grading rules are paired with automated checks.", "record_ids": [ "afterquery.product.round2-expert-signals", "afterquery.product.oct14-4" ], "claim_ids": [ "afterquery.claim.round2-product-expert-signals", "afterquery.claim.oct14-product-4" ], "aliases": [ "Rubric and Verifier-based RL" ] }, { "name": "Enterprise model services", "description": "Identifies gaps in a model’s abilities and prepares training data. Further training adapts a specialist model or agent for deployment.", "record_ids": [ "afterquery.product.round2-enterprise-model-work", "afterquery.product.oct14-5" ], "claim_ids": [ "afterquery.claim.round2-product-enterprise-model-work", "afterquery.claim.oct14-product-5" ], "aliases": [ "Custom Evals and Training Datasets" ] }, { "name": "Atlas and Raine Search", "description": "Atlas helps label and organize data. Raine Search finds past deals using everyday-language queries. Deployed for a named client; general self-service availability is undisclosed.", "record_ids": [ "afterquery.product.round2-atlas-raine-search" ], "claim_ids": [ "afterquery.claim.round2-product-atlas-raine-search" ] } ], "relationships": [ { "name": "NVIDIA", "relationship": "dataset_user", "description": "NVIDIA used AfterQuery tasks to train models for office work. Model-generated action records supported initial student-model training and later training with rewards.", "record_ids": [ "afterquery.customer.round2-0" ], "claim_ids": [ "afterquery.claim.round2-relationship-0" ] }, { "name": "Thinking Machines Lab", "relationship": "dataset_user", "description": "AfterQuery says it supplied training data for Inkling.", "record_ids": [ "afterquery.customer.round2-1" ], "claim_ids": [ "afterquery.claim.round2-relationship-1" ] }, { "name": "Thinking Machines Lab", "relationship": "technology_vendor", "description": "AfterQuery used Tinker for further terminal-agent training and office-task distillation (training one model with outputs from another model).", "record_ids": [ "afterquery.customer.round2-2" ], "claim_ids": [ "afterquery.claim.round2-relationship-2" ] }, { "name": "Legora", "relationship": "research_collaborator", "description": "Legora credits AfterQuery with co-creating a public synthetic BAR case. AfterQuery also reviewed how Legora prepared its data.", "record_ids": [ "afterquery.customer.round2-3" ], "claim_ids": [ "afterquery.claim.round2-relationship-3" ] }, { "name": "The Raine Group", "relationship": "implementation_partner", "description": "An investor and Raine Search deployment partner. AfterQuery worked with Raine bankers to search past deals using Atlas and deal categories specific to the firm.", "record_ids": [ "afterquery.customer.round2-4" ], "claim_ids": [ "afterquery.claim.round2-relationship-4" ] }, { "name": "Motif Technologies", "relationship": "dataset_user", "description": "AfterQuery identifies itself as the sole data partner for Motif 3.", "record_ids": [ "afterquery.customer.round2-5" ], "claim_ids": [ "afterquery.claim.round2-relationship-5" ] } ], "people": [ { "person_id": "afterquery.person.spencer-mateega", "role": "Co-founder · CEO", "bio": "Studied finance and statistics at Wharton, with computer science at Penn. His internships spanned finance and software engineering.", "claim_ids": [ "afterquery.claim.round2-person-spencer-mateega" ], "portrait": "assets/people/afterquery-spencer-mateega.jpg", "portrait_source_url": "https://www.ycombinator.com/companies/afterquery" }, { "person_id": "afterquery.person.carlos-georgescu", "role": "Co-founder · CTO", "bio": "Previously founded an education-technology business that was acquired. His software-engineering internships included Citadel Securities and Google.", "claim_ids": [ "afterquery.claim.round2-person-carlos-georgescu" ], "portrait": "assets/people/afterquery-carlos-georgescu.jpg", "portrait_source_url": "https://www.ycombinator.com/companies/afterquery" }, { "person_id": "afterquery.person.danny-tang", "role": "Early founding team · FinanceQA coauthor", "bio": "The launch describes his Wharton connection and an earlier startup with Spencer. Current title and employment status are unverified.", "claim_ids": [ "afterquery.claim.round2-person-danny-tang" ] } ], "notes": [ { "text": "Y Combinator’s current founder list names Spencer and Carlos. Launch material and the CEO’s account also name Danny Tang as an early team member. Danny’s current title and employment status are unverified.", "claim_ids": [ "afterquery.claim.round2-early-team-context" ] } ], "performance_chart": { "title": "NVIDIA’s GDPval warmup comparison", "subtitle": "Nemotron 3 Ultra · Table 4: initial training comparison", "series": [ { "label": "Without warmup", "value": 35.3 }, { "label": "With warmup", "value": 46.7 } ], "qualification": "The 11.4-point difference compares the warmup stage within NVIDIA’s training method. It does not isolate the effect of changing data suppliers.", "claim_ids": [ "afterquery.claim.round2-research-nvidia-warmup" ] }, "operating_case": { "title": "Legal evaluation with Legora", "description": "Legora’s full evaluation set has 5,161 cases and 11,075 documents across 28 practice areas. BAR uses a subset of hundreds. One co-created synthetic case is public.", "claim_ids": [ "afterquery.claim.round2-research-legora-bar" ] }, "combine_round_and_valuation": true, "capital_note": "The company says the round closed several months earlier. The exact closing date is undisclosed.", "metric_ids": [ "afterquery.metric.round2-financial-2", "afterquery.metric.oct14-2-expert-network" ], "metric_context": { "afterquery.metric.oct14-2-expert-network": { "scope": "Professional network", "note": "Undated company claim" } } } ] }