Rohit Patel leaves Meta to build Ndoc, an auditable AI-native rival to Office and Workspace

TechCrunch reports that Rohit Patel, who spent eight years building AI at Meta and was most recently a director at Meta Superintelligence Labs, is leaving to launch Ndoc, an AI-native alternative to Microsoft Office and Google Workspace. He joins a run of high-profile Meta departures including Joel Kaplan, Andrew Tulloch and Emily Dalton Smith. One exit is a personnel note; a string across policy, research and product starts to look like a pattern worth watching.

Ndoc is running an enterprise pilot and an alpha for a limited waitlist. Sign-up is free, with charges for usage and tokens. A beta is planned in the coming weeks and a full launch next year. Patel says it’s built and fine-tuned on a mix of open-source models and includes connectors to import data from existing tools.

Most enterprise AI is attached to existing software: it fills cells, drafts slides and summarizes documents. Patel compares that to putting lights on the bulls that till the soil when you should be building a tractor. The line lands, because much AI at work really is the old workflow with autocomplete. The incumbents are going the other way, from Microsoft’s Copilot super app (coverage) to Google’s Gemini agent for work (coverage) and the labs’ own office suites from Anthropic (coverage) and OpenAI (coverage).

Each Ndoc document pairs a presentation layer, a narrative layer with citations and a model layer that holds the data and Python scripts behind the numbers. If AI produces the decks executives use to make decisions, auditability is essential. A finance lead should be able to ask which line holds a formula and get a real answer. The big suites should copy this.

The open-format plan matters too. Patel says an open, exportable file that runs in any browser is part of why he left. Lock-in has defined office software for decades. A new format that breaks it would be welcome; one that becomes a new walled garden changes nothing.

This is a team of three with no funding against OpenAI, Anthropic, Google, Meta and others, all wiring models into Workspace and Microsoft 365, with OpenAI’s Space putting a shared workspace directly inside ChatGPT. Habits are sticky and incumbents own distribution. Patel’s early users told him they cared less about lower cost from open-source models than about having everything end to end in one place. That is a product insight and a warning: if the moat is experience, incumbents can imitate it more easily than it seems.

The speed-first philosophy deserves some caution. Moving fast on an imperfect plan is fine for a startup. In a tool that generates the numbers companies act on, errors compound quietly. Verification has to be as fast as generation, or transparency becomes decoration. More in our startups coverage, Meta coverage and Microsoft coverage.