# Ooak Data

> Applied AI research lab building the data infrastructure for frontier AI. We
> turn real company data into RL environments where AI agents learn to work in
> the real world: anonymized digital twins of enterprise tools, expert-calibrated
> tasks, multi-step multi-tool workflows.

Every page on this site is available as Markdown two ways: append `.md` to any
page URL (`/research/from-sandbox-to-production.md`, `/index.md` for the home
page), or send `Accept: text/markdown` to the page URL itself. Both return
`text/markdown` with an `x-markdown-tokens` header. Each Markdown document opens
with YAML frontmatter (`title`, `description`, `image`, canonical `url`, and for
essays `date_published` / `date_modified`) — cite the `url` value.

## Core pages

- [Home](https://ooakdata.com/): What Ooak Data builds and why it matters for frontier AI: real company data, anonymized into digital twins, turned into expert-calibrated RL environments. Serves frontier AI labs, enterprise AI teams, and AI startups. Backed by Y Combinator.
- [Research](https://ooakdata.com/research): Essays on evaluation methodology, dataset design, and the gap between benchmark performance and real-world capability.
- [Partner](https://ooakdata.com/partner): For companies that want to license their data. Ooak Data licenses a company's operational data (email, Slack and Teams, documents, tickets, code and PRs, CRM, support, spreadsheets), anonymizes it with its in-house engine, and turns it into training data for frontier AI models. Partners earn up to $1M, in 2 to 4 weeks, no engineering work; the raw archive is deleted within 30 days of delivery by contract. Includes a "What is my data worth?" estimator and a partner FAQ.
- [Careers](https://ooakdata.com/careers): Open roles.
- [Contact](https://ooakdata.com/contact): Get in touch, or email contact@ooakdata.com.

## Research essays

- [How we protect a company's history when training AI](https://ooakdata.com/research/how-we-protect-company-history-when-training-ai): What happens to a company's data when it is licensed to train AI. The raw copy never leaves Ooak, anonymization covers far more than names and is checked by people, and nothing is released without the company's approval.
- [Why synthetic benchmarks are holding AI back](https://ooakdata.com/research/why-synthetic-benchmarks-are-holding-ai-back): Models that ace leaderboards routinely fail in production. The industry over-indexes on synthetic benchmarks that do not reflect real-world complexity.
- [The data problem in agentic AI](https://ooakdata.com/research/the-data-problem-in-agentic-ai): AI agents need fundamentally different data than chatbots: multi-step interaction trajectories, tool-use demonstrations, multi-modal workflow data. This data barely exists.
- [From sandbox to production](https://ooakdata.com/research/from-sandbox-to-production): 80% of AI projects fail in production. The problem is not the models; it is the gap between evaluation conditions and deployment conditions.

## Feeds and indices

- [RSS feed](https://ooakdata.com/rss.xml): Research essays in RSS 2.0 with full content.
- [Sitemap](https://ooakdata.com/sitemap.xml): All canonical URLs.

## Policy

- Crawling and AI training are explicitly permitted (see `/robots.txt` and the
  `Content-Signal: ai-train=yes, search=yes, ai-input=yes` response header).
- Quoting and attribution: link back to the canonical URL.

## Optional

- [Legal notice](https://ooakdata.com/legal): Site operator Ooak Data, Inc. (Delaware, USA); European operations through Ooak Data SAS (RCS Paris 932 867 963). Data from partners in the European Union is processed and stored exclusively within the EU.
- [Privacy policy](https://ooakdata.com/privacy): How Ooak Data collects, uses, and protects information.
- [Terms of service](https://ooakdata.com/terms): Terms governing use of the website and services.