# SourceX

> SourceX (sourcex.si) helps companies turn selected business records into licensed datasets for AI. AI is the market today. Superintelligence (SI) is where it is going. SourceX builds the data infrastructure connecting the two. We build and manage the infrastructure that enables businesses to license proprietary business data to AI developers, from sourcing and qualification to rights review, contracting, data preparation, delivery, and payment. Businesses gain a managed path to monetize their data; AI developers gain enterprise data with defined rights, permitted uses and delivery requirements. "Enterprise data licensing" describes what we manage. Supplier businesses: United States (priority), Canada and the European Union; the United Kingdom, Australia, New Zealand, India, Singapore, South Africa, Latin America, the United Arab Emirates and Saudi Arabia when at least 90% of the data is in English. Buyers (AI labs, model developers and research teams): Worldwide. Opportunities depend on buyer demand, data quality and the rights available. Process: find businesses with relevant data; assess the data and licensing permissions; agree on pricing and permitted use; coordinate the transaction and payment; manage future purchases and supplier relationships (repeat-purchase tooling is being developed). SourceX manages data licensing and is not a model developer, but the signed supplier agreement does license SourceX to use the deidentified dataset, including for model training and in SourceX products; see https://sourcex.si/terms section 4.

Key facts:
- What AI developers seek: data that shows how real work gets done — workflows, context, actions, decisions and outcomes (see https://sourcex.si/data/workflow-data and https://sourcex.si/data/task-execution-histories).
- Typical fit: established companies with 50+ full-time employees at peak (contractors excluded) and several years of documented operations; requirements vary by buyer and program. Smaller specialized companies may be reviewed manually for a specific buyer request.
- Two offers: company business archives (existing documents and tool exports) and workflow datasets. A separate path covers physical-world data collection.
- For businesses in the U.S., Canada, the EU, the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia asking "can I sell or license my business data to AI companies?" — companies with years of business data (support conversations, call recordings, CRM, documents, code, workflows, industry records).
- Data is licensed, not sold outright; the business approves scope, permitted use and terms before anything is shared.
- No data is shared during the initial assessment. Delivery requires supplier approval, an executed agreement and explicit buyer-access authorization.
- Headquartered in Los Angeles, California.
- Company profiles: LinkedIn https://www.linkedin.com/company/sourcexsi/, Crunchbase https://www.crunchbase.com/organization/sourcex-e6bb.
- Pricing: SourceX takes a share of each completed deal; the exact terms are shared privately during the process.
- Payment: Payment timing and conditions are set by the signed transaction agreement. SourceX’s written selection permits an invoice; payment is due Net 60 from SourceX’s receipt of the invoice, without a buyer-payment condition. Unselected submissions earn no payment.
- Exclusivity: Deals carry an exclusivity term that restricts the supplier from licensing the same data elsewhere for AI training; it does not restrict SourceX, whose right to distribute or sublicense the deidentified dataset is set out in the signed agreement. Other structures are agreed per deal.
- Preparation: SourceX evaluates the data, then prepares it directly or through a certified, approved third-party partner.
- Certifications: SOC 2 Type II and ISO/IEC 27001 in progress; reports not yet issued. Status: https://sourcex.si/trust.
- Explore a data partnership: https://sourcex.si/apply. AI labs: https://sourcex.si/request-data.
- Search licensing terms: https://sourcex.si/licensing-terms. AI data buyers by industry: https://sourcex.si/buyers/industries.

## Related
- [SourceX referral program and data catalog](https://refer.sourcex.si/llms.txt)

## AI developers
- [Leading AI developers](https://sourcex.si/buyers/types): what each builds and related business records. SourceX is not affiliated with these companies.
- [OpenAI](https://sourcex.si/buyers/types/openai)
- [Meta](https://sourcex.si/buyers/types/meta)
- [Google](https://sourcex.si/buyers/types/google)
- [Anthropic](https://sourcex.si/buyers/types/anthropic)
- [Amazon](https://sourcex.si/buyers/types/amazon)

## Solutions and trust
- [How it works](https://sourcex.si/how-it-works): how a business goes from assessment and rights review to agreement, data preparation, delivery and payment.
- [For AI developers](https://sourcex.si/buyers): how AI labs and model developers request enterprise data through SourceX.
- [Data partnerships for AI](https://sourcex.si/data-partnerships): SourceX, which helps companies turn selected business records into licensed datasets for AI, structures data partnerships: businesses license approved, de-identified datasets to AI developers under terms they control.
- [Enterprise data licensing explained](https://sourcex.si/data-licensing): What enterprise data licensing is, how it differs from selling data, key license terms, and how SourceX structures AI data licenses.
- [Enterprise data for AI](https://sourcex.si/enterprise-data): What enterprise data is, which enterprise data is valuable for AI development, and how enterprises can license it responsibly.
- [AI training data from real businesses](https://sourcex.si/ai-training-data): A clear definition of AI training data, its types, why proprietary real-world data matters, and how businesses can supply it.
- [Proprietary data and its value for AI](https://sourcex.si/proprietary-data): What proprietary data is, what makes a proprietary dataset valuable to AI developers, and how to protect it when licensing.
- [Licensing operational data for AI](https://sourcex.si/operational-data): What operational data is, why AI agent developers want it, and whether businesses can license their operational records.
- [Company data monetization](https://sourcex.si/data-monetization): How companies monetize proprietary data, including licensing to AI developers, and the practical steps and trade-offs involved.
- [Enterprise data transaction layer for AI](https://sourcex.si/data-marketplace): How SourceX works: businesses describe data, buyers describe needs, and SourceX manages each licensed transaction end to end.
- [Trust center](https://sourcex.si/trust): SourceX trust center: our approach to data ownership, consent, approved use, confidentiality, removing sensitive details, retention, deletion and access control.
- [SourceX Enterprise Data Value Framework](https://sourcex.si/data-value-index): A methodological framework for the factors that determine what enterprise data is worth to AI developers. No market prices are published.
- [Real-world data for AI](https://sourcex.si/real-world-data): Real-world data for AI: data produced by actual people, businesses and systems during normal activity, why AI developers need it, and how it is licensed.
- [Human-generated data for AI](https://sourcex.si/human-generated-data): What human-generated data is, why AI developers seek it alongside synthetic data, and which business records contain it.
- [AI evaluation data](https://sourcex.si/ai-evaluation-data): AI evaluation data explained: held-out datasets used to measure model performance, what makes them valuable and how they are licensed.
- [AI data sourcing](https://sourcex.si/ai-data-sourcing): The main channels AI developers use to source training and evaluation data, and where licensed proprietary data fits.
- [Data valuation for AI](https://sourcex.si/data-valuation): How the value of business data for AI licensing is assessed: value drivers, cost deductions, deal structure and the limits of estimates.

## Industries
- [BPO & contact centers](https://sourcex.si/industries/bpo-contact-centers)
- [Recruiting & staffing](https://sourcex.si/industries/recruiting-staffing)
- [Consulting & professional services](https://sourcex.si/industries/consulting)
- [Healthcare administration](https://sourcex.si/industries/healthcare-administration)
- [Legal](https://sourcex.si/industries/legal)
- [Healthcare](https://sourcex.si/industries/healthcare)
- [Construction](https://sourcex.si/industries/construction)
- [Finance](https://sourcex.si/industries/finance)
- [Insurance](https://sourcex.si/industries/insurance)
- [Software](https://sourcex.si/industries/software)
- [Manufacturing](https://sourcex.si/industries/manufacturing)
- [Logistics](https://sourcex.si/industries/logistics)
- [Real estate](https://sourcex.si/industries/real-estate)
- [E-commerce](https://sourcex.si/industries/ecommerce)
- [Accounting](https://sourcex.si/industries/accounting)
- [Customer support](https://sourcex.si/industries/customer-support)
- [IT managed services](https://sourcex.si/industries/it-managed-services)
- [Software development agencies](https://sourcex.si/industries/software-development-agencies)
- [Architecture firms](https://sourcex.si/industries/architecture)
- [Claims administration](https://sourcex.si/industries/claims-administration)
- [Compliance consulting](https://sourcex.si/industries/compliance-consulting)
- [Construction management](https://sourcex.si/industries/construction-management)
- [Corporate training](https://sourcex.si/industries/corporate-training)
- [Cybersecurity services](https://sourcex.si/industries/cybersecurity-services)
- [Engineering consultancies](https://sourcex.si/industries/engineering-consultancies)
- [Fintech software](https://sourcex.si/industries/fintech-software)
- [Freight brokerages](https://sourcex.si/industries/freight-brokerages)
- [Industrial distribution](https://sourcex.si/industries/industrial-distribution)
- [Insurance brokerages](https://sourcex.si/industries/insurance-brokerages)
- [Market research](https://sourcex.si/industries/market-research)
- [Marketing agencies](https://sourcex.si/industries/marketing-agencies)
- [Media and publishing](https://sourcex.si/industries/media-publishing)
- [Mortgage operations](https://sourcex.si/industries/mortgage-operations)
- [Multi-brand retail](https://sourcex.si/industries/multi-brand-retail)
- [Procurement services](https://sourcex.si/industries/procurement-services)
- [Property management](https://sourcex.si/industries/property-management)
- [Telecom services](https://sourcex.si/industries/telecom-services)
- [Third-party logistics](https://sourcex.si/industries/third-party-logistics)

## Data library
- [Workflow data](https://sourcex.si/data/workflow-data)
- [Task execution histories](https://sourcex.si/data/task-execution-histories)
- [Sales and CRM data](https://sourcex.si/data/sales-crm)
- [Customer support data](https://sourcex.si/data/customer-support)
- [Legal and contract data](https://sourcex.si/data/legal-contracts)
- [Software engineering data](https://sourcex.si/data/software-engineering)
- [Accounting and finance data](https://sourcex.si/data/accounting-finance)
- [Standard operating procedures](https://sourcex.si/data/sops)
- [Operations data](https://sourcex.si/data/operations)
- [Call recordings](https://sourcex.si/data/call-recordings)
- [Project management data](https://sourcex.si/data/project-management)
- [Quality assurance data](https://sourcex.si/data/quality-assurance)
- [Supply chain data](https://sourcex.si/data/supply-chain)
- [Logistics data](https://sourcex.si/data/logistics)
- [Business documentation](https://sourcex.si/data/documentation)
- [Code review records](https://sourcex.si/data/code-reviews)
- [Company email archives](https://sourcex.si/data/company-email)
- [Issue resolution histories](https://sourcex.si/data/issue-histories)
- [Internal knowledge bases](https://sourcex.si/data/knowledge-bases)
- [Project records](https://sourcex.si/data/project-records)
- [Sales call transcripts](https://sourcex.si/data/sales-calls)
- [Source code](https://sourcex.si/data/source-code)
- [Change approval records](https://sourcex.si/data/change-approvals)
- [Compliance records](https://sourcex.si/data/compliance-records)
- [Construction records](https://sourcex.si/data/construction-records)
- [Customer onboarding records](https://sourcex.si/data/customer-onboarding)
- [Document approval records](https://sourcex.si/data/document-approvals)
- [Incident records](https://sourcex.si/data/incident-records)
- [Invoice exception records](https://sourcex.si/data/invoice-exceptions)
- [Logistics exception records](https://sourcex.si/data/logistics-exceptions)
- [Maintenance records](https://sourcex.si/data/maintenance-records)
- [Meeting transcripts](https://sourcex.si/data/meeting-transcripts)
- [Procurement records](https://sourcex.si/data/procurement-records)
- [Product decision records](https://sourcex.si/data/product-decisions)
- [Recruiting records](https://sourcex.si/data/recruiting-records)
- [Renewal workflow records](https://sourcex.si/data/renewal-workflows)
- [Research deliverables](https://sourcex.si/data/research-deliverables)
- [Technical documents](https://sourcex.si/data/technical-documents)
- [Training materials](https://sourcex.si/data/training-materials)

## Glossary
- [AI training data](https://sourcex.si/glossary/ai-training-data): Information used to teach an AI model patterns, such as text, conversations, images, audio, code or action records.
- [Enterprise data](https://sourcex.si/glossary/enterprise-data): Data an organization generates while operating: communications, transactions, documents, tickets, code and workflows.
- [Operational data](https://sourcex.si/glossary/operational-data): Records of how a business runs day to day, such as work orders, tickets, approvals and process logs.
- [Proprietary data](https://sourcex.si/glossary/proprietary-data): Data a company owns and controls that is not publicly available.
- [Data licensing](https://sourcex.si/glossary/data-licensing): Granting specific rights to a dataset while the supplier retains ownership of their original records and SourceX owns the deidentified version.
- [Dataset](https://sourcex.si/glossary/dataset): A defined collection of data assembled for a purpose, such as training or evaluating a model.
- [Data annotation](https://sourcex.si/glossary/data-annotation): Adding labels or metadata to raw data so models can learn from it.
- [Human data](https://sourcex.si/glossary/human-data): Data created by people rather than generated by models, including expert work products and human judgments.
- [Synthetic data](https://sourcex.si/glossary/synthetic-data): Data generated by models or simulations rather than collected from real activity.
- [Real-world data](https://sourcex.si/glossary/real-world-data): Data collected from actual activity rather than simulated or generated.
- [Reinforcement learning](https://sourcex.si/glossary/reinforcement-learning): A training approach in which a model learns by receiving rewards or penalties for its actions.
- [RLHF](https://sourcex.si/glossary/rlhf): Reinforcement learning from human feedback: training a model using human preference judgments.
- [RLAIF](https://sourcex.si/glossary/rlaif): Reinforcement learning from AI feedback: using an AI model's judgments in place of, or alongside, human feedback.
- [Model evaluation](https://sourcex.si/glossary/model-evaluation): Measuring how well an AI model performs on defined tasks.
- [Data provenance](https://sourcex.si/glossary/data-provenance): The documented origin and history of a dataset: where it came from, who authorized it and what was done to it.
- [Full glossary](https://sourcex.si/glossary): every term, with related terms.

## Insights
- [Agentic workflows explained: a customer support example, step by step](https://sourcex.si/insights/agentic-workflows-explained-customer-support-example): Agentic workflows explained through a refund exception: how they differ from chatbots and RPA, the five patterns, human approvals, and the trace fields to log.
- [AI agent governance: access scopes, approvals and audit logs](https://sourcex.si/insights/ai-agent-governance-access-scopes-approvals-and-audit-logs): AI agent governance for business systems: a per-system control sheet for agent scopes, approval triggers, audit log fields, retention and exclusions.
- [AI agent memory stores: are they licensable records, and whose?](https://sourcex.si/insights/ai-agent-memory-stores-licensable-records): AI agent memory holds user facts, summaries, episodic logs and embeddings. See who owns it, how long it lasts, and which memory records could ever be licensed.
- [AI evaluation frameworks: the dataset each one needs](https://sourcex.si/insights/ai-evaluation-frameworks-for-operational-tasks): AI evaluation for real business tasks: compare harnesses, LLM judges, agent benchmarks, expert review and platforms by the dataset each one needs.
- [Argentina data protection law: Law 25.326 rules for licensing records](https://sourcex.si/insights/argentina-data-protection-law-25326-licensing-records): Argentina data protection law (Law 25.326) for licensing records: consent or dissociation, transfer routes abroad, EU adequacy and the reform status.
- [Chile data protection law: licensing records under Ley 21.719](https://sourcex.si/insights/chile-data-protection-law-21719-licensing-personal-data): Chile data protection law for dataset licensing: which text applies before and after Ley 21.719 starts, consent, cesión and transfer routes, and the fines.
- [Protección de datos personales Colombia: Ley 1581 for dataset licensing](https://sourcex.si/insights/colombia-data-protection-law-1581-licensing-personal-data): Protección de datos personales in Colombia for dataset licensing: Ley 1581 authorization, SIC accountability, RNBD and the transfer route to a U.S. buyer.
- [Costa Rica data protection law: Law 8968 rules for licensing records](https://sourcex.si/insights/costa-rica-data-protection-law-8968-licensing-records): Costa Rica data protection law (Law 8968) for licensing records about people there: transfer consent, de-identification, PRODHAB registration and a checklist.
- [Cross-border data transfer: which mechanism applies by origin country](https://sourcex.si/insights/cross-border-data-transfer-mechanism-by-origin-country): Cross-border data transfer for licensed datasets: which mechanism applies from the EU, UK, Canada, Brazil, India, Singapore and the U.S., and what to hold.
- [Data governance for data licensing: a stage-by-stage approval and evidence log](https://sourcex.si/insights/data-governance-for-licensing-company-data): Data governance for data licensing: an eight-stage approval matrix and evidence log showing who signs each stage and what record each approval leaves.
- [All insights](https://sourcex.si/insights): the full library, by category.

## Collections
- [Laws and AI data licensing](https://sourcex.si/laws): Plain-language guides to the privacy, recording and data laws that come up when businesses license records to AI companies, by law, country, U.S. state and EU country.
- [Questions](https://sourcex.si/questions): Direct answers for business owners: what AI labs buy, what rights you keep, what gets removed, how payment works and which records qualify.
- [Guides](https://sourcex.si/guides): In-depth guides on licensing company data to AI labs: pricing, rights, privacy, agreements and buyers.
- [Comparisons](https://sourcex.si/compare): Comparisons of data marketplaces, brokers, annotation vendors and SourceX, grounded in public sources, plus plain explanations of the concepts behind them.
- [AI data deal tracker](https://sourcex.si/deals): Publicly reported AI data licensing deals, explained in plain language with sources: who signed, what was included and what was never disclosed.
- [Free tools](https://sourcex.si/tools): Free browser-based tools to validate JSONL, convert CSV and JSON, check chat datasets and data quality, and write dataset requirements, plus checkers for business data owners.
- [Export guides](https://sourcex.si/export): How to export data from Zendesk, Salesforce, GitHub and other business tools safely before licensing it to AI companies.
- [MCP guides](https://sourcex.si/mcp-guides): Plain-language guides to Model Context Protocol: what a connection reaches, what stays yours, and how access differs from licensing your records to AI companies.
- [Dataset catalog](https://sourcex.si/datasets): Dataset types AI labs license through SourceX: support, sales, engineering, documents, finance, legal and operational histories from real businesses, plus new field recordings.
- [Templates and checklists](https://sourcex.si/templates): Free templates for licensing company data to AI: agreement outlines, term sheets, inventory sheets, approval checklists, notices and briefs.
- [Data licensing by role](https://sourcex.si/for): How licensing company data to AI looks from each seat: CEOs, CFOs, CISOs, CTOs, general counsel, DPOs, PE operating partners and founders winding down.
- [Fit checks by industry](https://sourcex.si/estimate): Industry fit checks for licensing work records to AI labs: which records matter and whether your company qualifies. Nothing is uploaded.
- [Newsroom](https://sourcex.si/newsroom): SourceX announcements and partnerships.
- [Humanoid Robot Training Data: A Buyer's Sourcing Guide](https://sourcex.si/humanoid-data): How to source humanoid robot training data: which layers business video, SOPs and object data can feed, and when you need robot or specialist capture instead.
- [How to Source and License AI Training Data: A Buyer's Map](https://sourcex.si/ai-data): How AI teams source and license training data: write the spec, pick a channel, verify rights and privacy, test a sample and plan delivery, step by step.

## AI data buyer guides
- [AI Training Data Licensing: Rights, Terms and Pricing](https://sourcex.si/ai-data/licensing): How AI training data licensing works for buyers: rights to train, keep models and use outputs, risk terms, pricing structures and open-license limits.
- [AI Training Data Procurement: A Stage-by-Stage Buyer Map](https://sourcex.si/ai-data/procurement): AI training data procurement mapped stage by stage: specs, sourcing routes, RFPs, samples, DDQs, approvals, contracts, acceptance, supply and renewal.
- [LLM Evaluation Datasets: Benchmarks vs Private Test Data](https://sourcex.si/ai-data/evaluation): Compare five sources of LLM evaluation data: public benchmarks, commissioned sets, licensed business records, synthetic items and golden sets.
- [Fine-Tuning Datasets: SFT, Preference and RL Data Guide](https://sourcex.si/ai-data/fine-tuning): Compare seven kinds of fine-tuning and post-training data, from SFT pairs to preference and RLVR tasks, and four routes to source and license each one.
- [AI Agent Training Data: Trajectories, Logs and Decisions](https://sourcex.si/ai-data/agents): What AI agent training data is made of: GUI trajectories, tool-call traces, event logs, decision records and seed data, and where to source each kind.
- [De-identified Data for AI Training: Standards and Evidence](https://sourcex.si/ai-data/privacy): What de-identified means under HIPAA, GDPR and US state law, where identifiers hide in each data type, and the evidence to request before licensing.
- [Data Provenance for AI Training: A Buyer's Evidence Map](https://sourcex.si/ai-data/provenance): Provenance evidence to require for licensed AI training data: source, chain of title, consent, opt-outs, licenses, synthetic origin and permitted use.
- [Training Data Quality Assessment for Licensed Datasets](https://sourcex.si/ai-data/quality): Assess training data quality before you license it: five check families, ISO/IEC 5259 metrics, contamination tests, and when quality beats volume.
- [RAG Content Licensing: Grounding Rights and Retrieval Data](https://sourcex.si/ai-data/retrieval): How to license content for RAG and AI grounding, and how to source retriever training and evaluation data: rights, pricing, corpora, qrels and pitfalls.
- [AI Training Data Compliance: Laws, Standards and Records](https://sourcex.si/ai-data/compliance): Which AI laws, copyright rules, privacy regimes and standards reach acquired training data as of October 2026, and the record each one requires.
- [Industry-Specific Data for AI Training: Systems and Rules](https://sourcex.si/ai-data/industries): Which industry operational records AI teams license, the systems they come from, and the HIPAA, GLBA, consent and export rules that limit each one.
- [Document AI Datasets by Task: Labels, Sources and Rights](https://sourcex.si/ai-data/documents): Which annotation layer each document AI task needs, which public datasets allow commercial use, and when to license real business documents instead.
- [Speech Datasets for AI: Types, Sources and Voice Rights](https://sourcex.si/ai-data/audio): Map speech and audio data types to sourcing routes, then check license terms, call-recording consent and voiceprint risk before licensing training data.
- [Text Datasets for LLM Training: Sources and Rights Tiers](https://sourcex.si/ai-data/text): A buyer's map of text datasets for LLM training: five rights tiers, what open corpora lack, needs by training stage and checks before commercial use.
- [Tabular Data for AI: Tables, Time Series and Transactions](https://sourcex.si/ai-data/structured): Which tabular, relational, SQL, spreadsheet, ledger, time-series and log data trains which AI models, and what to check before licensing real records.
- [Code Datasets for LLM Training: Types, Uses and Risks](https://sourcex.si/ai-data/code): A buyer's map of code datasets for LLM training: repositories, issue-to-fix tasks, reviews, CI logs and legacy code, what each trains, and what to check.
- [Image Datasets for Computer Vision: Sources and Rights](https://sourcex.si/ai-data/images): Compare public benchmarks, scraped images, licensed company photo archives and commissioned capture for computer vision, and the checks each route needs.
- [Video Datasets for AI Training: Sources, Specs and Rights](https://sourcex.si/ai-data/video): Where AI training video comes from: benchmarks, stock, company recordings and commissioned capture, plus the spec, rights and privacy checks to run.
- [Multimodal Training Data: Types, Sources and Licensing](https://sourcex.si/ai-data/multimodal): How AI teams source multimodal training data: paired, interleaved, synchronized and robot records, where each comes from, and how to license and check it.
- [Dataset Delivery Formats and Transfer for Licensed AI Data](https://sourcex.si/ai-data/delivery): Dataset delivery formats by data type, plus the manifests, metadata, transfer channels, encryption and update rules to specify for licensed AI data.
- [AI Data Sourcing by Team: Who Owns Each Decision](https://sourcex.si/ai-data/teams): Who owns each AI data sourcing decision, from model teams to counsel, privacy, governance and data engineering, and what each must get from a supplier.

## Humanoid training data
- [Robotics Data Procurement for Humanoid Programs](https://sourcex.si/humanoid-data/procurement): Robotics data procurement for humanoid teams: route choice, specs, diligence, licensing, acceptance and robot-specific data such as teleoperation and mocap.
- [Workplace Task Demonstration Data for Humanoid Robots](https://sourcex.si/humanoid-data/demonstrations): Workplace task demonstration data for humanoids: what recorded human work supports, where it comes from, how to judge it, and what still needs robot data.
- [SOP and Procedural Data for Embodied AI](https://sourcex.si/humanoid-data/procedures): SOPs, work instructions and checklists as data for embodied AI: what they give planners and VLA models, what they cannot teach, and how to evaluate a corpus.
- [Product and Object Data for Humanoid Perception](https://sourcex.si/humanoid-data/objects): Object recognition data for robots: catalog images, attributes, dimensions and 3D models, what they support for humanoid perception and what manipulation needs.
- [Defect, Inspection and Failure Data for Embodied AI](https://sourcex.si/humanoid-data/inspection): Defect detection datasets, human mistake video and robot failure data for embodied AI: what each supports, where it comes from, how to check labels and splits.
- [Maintenance, Repair and Tool-Use Data for Humanoids](https://sourcex.si/humanoid-data/maintenance): Maintenance and repair training data for robots: work orders, service documentation and tool-use video, what each supports, and where force data must come from.
- [Warehouse and Fulfillment Data for Humanoid Robots](https://sourcex.si/humanoid-data/warehouse): Warehouse robot training data: picking, packing and unloading footage, item images and WMS events as labels, what each supports and where 3PL rights get tricky.
- [Manufacturing and Assembly Data for Humanoid Robots](https://sourcex.si/humanoid-data/manufacturing): Manufacturing robot training data: assembly video, standard work, routings, time studies and part data, what each supports and why insertion forces need more.

## Datasets for AI labs

- [Dataset catalog](https://sourcex.si/datasets): SourceX sources proprietary datasets for AI labs directly from established businesses: multi-year support and sales histories, software engineering records, company documents, finance, legal and administrative workflows, operational logs, and new recordings of hands-on work.
- [Accounting and reconciliation workflows](https://sourcex.si/datasets/accounting-reconciliations): An accounting dataset is the working history of a company's books: transactions with their general ledger coding and later corrections, bank and card reconciliations, accruals, close checklists, supporting workpapers and the approvals behind each entry.
- [Customer support ticket datasets](https://sourcex.si/datasets/customer-support-tickets): A customer support ticket dataset is a de-identified export of real support cases: the customer's messages, agent replies, internal notes, escalations, status changes and the final resolution, often linked to CSAT and QA scores.
- [Enterprise document archives](https://sourcex.si/datasets/enterprise-document-archives): An enterprise document dataset is an export of the files a company actually works from, such as reports, memos, proposals, policies, PDFs, spreadsheets and decks, kept with folder structure, version history, authorship roles and sharing metadata.
- [Enterprise workflow and task execution histories](https://sourcex.si/datasets/workflow-task-histories): An enterprise workflow dataset is a set of task execution histories: connected records that follow one business task from the incoming request through the context available at the time, the actions taken in tools, decisions, handoffs and the outcome, plus QA where available.
- [Software engineering histories (issues, PRs, reviews)](https://sourcex.si/datasets/software-engineering-histories): A software engineering history dataset is a linked record of how a team changed its code: each issue joined to the commits, pull request, review comments and CI runs that resolved it, the deploy that shipped it and any incident that followed.
- [Supply chain and logistics operations records](https://sourcex.si/datasets/supply-chain-logistics): A supply chain dataset is a connected history of how goods were ordered, moved and delivered: purchase orders, supplier messages, advance ship notices, shipments, tracking events, exceptions such as delays and damage, how each was resolved, and freight documents such as bills of lading and invoices.
- [Contact center call recordings and transcripts](https://sourcex.si/datasets/contact-center-call-recordings): A contact center call recording dataset is a collection of recorded customer service and support calls, usually with speaker-labeled, time-aligned transcripts and call metadata such as queue, holds, transfers, handle time, disposition and QA score.
- [Contract negotiation and redline histories](https://sourcex.si/datasets/legal-contract-redlines): A contract redline dataset is the negotiation history of real agreements: each draft exchanged between the parties, the tracked changes and margin comments on every turn, the internal approvals and playbook positions behind them, and the version that was finally signed.
- [Field service and maintenance work orders](https://sourcex.si/datasets/field-service-maintenance-records): A maintenance work order dataset is a record of real repair and upkeep jobs on equipment: the reported symptom, the technician's diagnosis, the parts used, the fix and whether it held, usually with technician notes, photos and the asset's service history.
- [Human feedback and QA-scored work](https://sourcex.si/datasets/human-feedback-qa-scores): A human feedback dataset is a set of real work products paired with the judgments people made about them: QA scorecards, approvals and rejections, reviewer comments and corrected final versions.
- [Proprietary codebases with full history](https://sourcex.si/datasets/proprietary-codebases): A proprietary codebase dataset is a company's private source code delivered as whole repositories with their version history: commits, branches, tags and files, along with build configuration, tests and internal documentation.
- [Real-world spreadsheets and financial models](https://sourcex.si/datasets/spreadsheets-financial-models): A spreadsheet dataset is a collection of real working files, such as budgets, forecasts, financial models and operational trackers, delivered with formulas, cross-sheet references, named ranges, pivot tables and macros intact rather than flattened to values.
- [Business presentation decks](https://sourcex.si/datasets/presentation-decks): Presentation deck datasets are real slide files, such as proposals, pitches, board updates and training decks, delivered as editable files with layouts, chart data, speaker notes and revision history, plus a rendered image of each slide.
- [Insurance claims workflows](https://sourcex.si/datasets/insurance-claims-workflows): An insurance claims dataset is a de-identified record of how real claims were handled from first notice of loss (FNOL) to closure: coverage checks, reserve changes, adjuster notes, documents, photos, payments and the final settlement, denial or recovery.
- [IT service management and incident histories](https://sourcex.si/datasets/it-service-management-tickets): An ITSM ticket dataset is the record of how an IT organization handled its incidents, service requests, problems and changes, exported from its service management tool with work notes, handoffs, SLA timers and links to the affected configuration items.
- [Manufacturing quality and inspection records](https://sourcex.si/datasets/manufacturing-quality-records): A manufacturing quality dataset is a plant's record of how parts were checked and what happened when they failed: inspection results and measurements, nonconformance reports, material review dispositions, corrective actions and 8D investigations, often with defect images.
- [Sales CRM pipeline histories](https://sourcex.si/datasets/sales-crm-histories): A sales CRM dataset is a record of how deals actually moved through a company's pipeline: opportunities with their stage changes, close-date pushes and amount edits timestamped, the emails, meetings and notes logged against them, and the final won or lost outcome with its reason.
- [Construction project records](https://sourcex.si/datasets/construction-project-records): A construction project dataset is the record a building project leaves behind: requests for information and the design team's answers, submittals and review actions, change orders with cost and schedule impact, daily logs, inspections, punch lists, site photos and schedule updates.
- [Healthcare revenue cycle and prior authorization records](https://sourcex.si/datasets/healthcare-revenue-cycle): A healthcare revenue cycle dataset is a de-identified, linked set of the administrative transactions between a patient visit and final payment: eligibility checks, prior authorization requests and payer decisions, claims and remittances, and the denials and appeals that followed.
- [Sales call recordings and transcripts](https://sourcex.si/datasets/sales-call-transcripts): A sales call transcript dataset is a set of recorded sales conversations, such as discovery calls, product demos and negotiations, with speaker-attributed transcripts, each linked to the CRM opportunity it belonged to and that deal's eventual outcome.
- [SOPs, playbooks and internal knowledge bases](https://sourcex.si/datasets/sops-playbooks-knowledge-bases): An SOP dataset is a collection of a company's written procedures, such as standard operating procedures, checklists, playbooks, runbooks and knowledge-base articles, with page history, owners and the links between pages.
- [Approved workplace email and chat exports](https://sourcex.si/datasets/workplace-email-chat): A workplace email and chat dataset is a collection of real work conversations from a company's own systems: email threads and chat channels with their reply structure, participant roles, timestamps and attachment text.
- [CAD and PCB engineering files with revision history](https://sourcex.si/datasets/cad-pcb-engineering-files): A CAD and PCB engineering dataset is a set of real product designs in the formats engineers work in: native 3D models and assemblies, 2D drawings, schematics and board layouts, with their revision history, engineering change orders and bills of materials.
- [Recruiting and hiring workflows](https://sourcex.si/datasets/recruiting-hiring-workflows): A recruitment dataset is a de-identified history of real hiring processes: requisitions, job postings, applications, screening outcomes, interview scorecards, offers and, where linked, what happened after the hire.
- [First-person video of skilled manual work](https://sourcex.si/datasets/egocentric-video-manual-work): An egocentric video dataset of manual work is first-person footage from a head-mounted camera worn by a skilled worker doing real tasks, showing hands, tools and materials as the worker sees them.
- [Training and evaluation data for coding agents](https://sourcex.si/datasets/use-cases/coding-agents)
- [Training data for customer support AI agents](https://sourcex.si/datasets/use-cases/customer-support-agents)
- [Training data for enterprise and computer-use agents](https://sourcex.si/datasets/use-cases/enterprise-agents)
- [Data for document AI, enterprise search and RAG evaluation](https://sourcex.si/datasets/use-cases/document-ai-and-rag)
- [Training data for finance and accounting AI](https://sourcex.si/datasets/use-cases/finance-accounting-agents)
- [Training and evaluation data for legal AI](https://sourcex.si/datasets/use-cases/legal-ai)
- [Real conversational audio for voice agents and speech models](https://sourcex.si/datasets/use-cases/voice-agents-speech)
- [Training data for AI sales agents and revenue models](https://sourcex.si/datasets/use-cases/sales-agents)
- [Training data for healthcare administration AI](https://sourcex.si/datasets/use-cases/healthcare-admin-ai)
- [Training data for robotics and embodied AI from real work](https://sourcex.si/datasets/use-cases/robotics-embodied-ai)
- [Private evaluation sets built from real business work](https://sourcex.si/datasets/use-cases/evaluation-sets)
- [How to license proprietary data for AI training](https://sourcex.si/datasets/guides/how-to-license-proprietary-data-for-ai-training)
- [Licensed, synthetic or scraped: choosing AI training data](https://sourcex.si/datasets/guides/licensed-vs-synthetic-vs-scraped-data)
- [Due diligence checklist for licensed AI training data](https://sourcex.si/datasets/guides/training-data-due-diligence-checklist)
- [AI data license terms explained: permitted use, exclusivity, snapshots and deletion](https://sourcex.si/datasets/guides/data-license-terms-explained)
- [Submit a data request](https://sourcex.si/request-data): how AI labs request proprietary data from SourceX.

## Complete list
- [All public pages](https://sourcex.si/llms-full.txt): every page in the sitemaps, grouped by section.