Rohit Patel has left Meta after eight years to launch Ndoc, a startup building an AI native workspace designed to replace Microsoft Office and Google Workspace. Patel was recently a director at Meta Superintelligence Labs, where he worked on reinforcement learning and on judging whether AI output can be trusted.
What Ndoc puts in one place
Ndoc brings together work that usually spreads across spreadsheets, documents, slide decks, dashboards and chatbot chats. Colleagues can open and use a shared workspace even if they have not signed up for Ndoc, which removes one common barrier to collaboration.
The workspace has three tabs. Stage holds presentations, slide decks, dashboards, simulations and small apps. Narrative holds the analysis behind the findings, with numbered citations and generated figures. Model holds the datasets and Python scripts behind the work.
Why checking the work matters
Patel argues that people should be able to inspect the code behind AI generated documents before they act on them. If AI is going to build the spreadsheets and slide decks that companies use to make decisions, he believes someone should be able to check its work.
“Even if you don’t know any Python, you can ask [the AI chat] to explain, ‘Hey, what does this script do? Which line has my formula?’ And you can verify that formula,” Patel said.
Users can upload a dataset and ask Ndoc to analyse it and write a narrative about the findings. They can also add context about the audience and tone, and Patel said the product keeps memory across the person, the document, the team and the company.
A file format of its own
Patel also plans to reinvent the document file itself. Ndoc documents can be exported as a single .ndoc.html file that opens in any browser, and he plans to open source the format so other developers can build compatible tools.
He said leaving Meta was partly about speed. Agreeing on an open format inside a large company would involve hundreds of people and opinions, while a small team can move faster on its own.
Ndoc is built and fine tuned using a mixture of open source models. Patel said building from scratch makes it considerably cheaper to run than alternatives built on frontier AI models.
Small team, crowded market
The bet is large for a team of three with no outside funding. Patel said he wants proof that people love the product before he seeks investment, and customers will pay for usage and tokens while sign up remains free.
Ndoc is running an enterprise pilot and opening an alpha version to a limited number of people on its waitlist. A beta version is planned in the coming weeks, with a full launch planned for next year.
The competition is formidable. OpenAI, Anthropic, Meta, xAI and Google have all built enterprise AI tools that connect with existing workplace apps, and OpenAI launched Space, a shared workspace inside ChatGPT, in September.
Patel also built connectors so customers can import data from the tools they already use. Changing work habits is hard, and that step may matter as much as any new feature in winning over skeptical teams.
