Rohit Patel is leaving Meta after eight years to build Ndoc, a startup he is pitching as an AI-native replacement for the Microsoft Office suite and Google Workspace. Patel, formerly a director at Meta Superintelligence Labs, spent his tenure on reinforcement learning—training systems through trial and error—and on checking whether AI output can be trusted.
His exit follows other senior departures from Meta, including Chief Global Affairs Officer Joel Kaplan, AI researcher Andrew Tulloch, and Head of Product for AI-for-Work Emily Dalton Smith. Patel is starting lean: a team of three and no outside funding so far, while large AI labs also push enterprise tools that plug into existing office software.
What Ndoc is trying to replace
Patel says Ndoc is meant to be more than another chatbot bolted onto spreadsheets and slides. The aim is one shared workspace instead of hopping across Excel, Word, PowerPoint, Tableau, coding tools and copy-paste from chatbots. Colleagues can open the work even if they are not signed up for Ndoc.
The product layout has three parts. The Stage tab holds presentations of findings—slide decks, dashboards, simulations or small apps. The Narrative tab carries the analysis, numbered citations, generated figures and the reasoning behind the numbers. The Model tab stores the datasets and Python scripts underneath. Users can upload data, ask for analysis and a written narrative, and set document context such as audience, tone and memory across a person, document, team and company.
That design tracks Patel’s Meta work on trustworthy AI. If models write the sheets and decks companies use for decisions, he argues, people should be able to inspect the work—for example by opening the Python that produced a document. Even without coding skills, he said, users can ask the chat to explain a script and verify a formula: “There’s nothing hidden behind the curtain here.”
Cost, format and competition
Patel says Ndoc is built and fine-tuned with a mix of open-source models and was developed from scratch, so he claims it costs less to run than products built on frontier models. Early trial users, he said, cared less about that cost story and more about seeing end-to-end work in one place.
He also wants to change the file itself. Documents can export as a single “.ndoc.html” file that opens in any browser, a format he plans to open source so other developers can build on it without the lock-in of long-standing office platforms. That open approach, he said, was part of why he left a large company where agreeing on a format would need “hundreds of people and opinions.”
OpenAI, Anthropic, Meta, SpaceXAI and Google have released enterprise AI that connects to Workspace, Microsoft 365 and similar apps. OpenAI also launched Space, a shared workspace inside ChatGPT. Patel called Space a possible threat but argued that merely filling spreadsheet cells with AI misses the point; he has added connectors so customers can import data from tools they already use.
Pilot, pricing and what comes next
As of Tuesday, Ndoc is in an enterprise pilot and opening a limited alpha to waitlist users. Sign-up is free; customers pay for usage and tokens. Patel hopes to ship a beta in the coming weeks and a full launch next year. He said he wanted proof that people love the product before raising money, and that moving fast on an imperfect plan matters more than waiting for a perfect one.
Aasan alfaaz mein
Meta ke AI director Rohit Patel company chor kar Ndoc naam ka naya startup bana rahe hain. Yeh tool Word, Excel aur Google Workspace jaisi cheezon ko AI se ek jagah laane ki koshish hai. Abhi limited pilot aur waitlist chal rahi hai; beta jaldi, poori launch agle saal. Office ya freelance kaam karne walon ke liye yeh naya option ho sakta hai, lekin pehle product ka asar dekhna hoga.
