Former Meta AI director Rohit Patel launches startup to rival Microsoft Office

Rohit Patel, who spent eight years developing artificial intelligence (AI) at Meta and served as a director at its Meta Superintelligence Labs unit, has left the company. He has founded a startup called Ndoc, which he describes as an AI-native replacement for the Microsoft Office and Google Workspace productivity suites, TechCrunch has reported. Patel is one of several high-profile figures to leave Meta recently, following Chief Global Affairs Officer Joel Kaplan, AI researcher Andrew Tulloch and AI-for-Work product head Emily Dalton Smith.

One workspace instead of many tools

Ndoc aims to replace the fragmented way people work across Excel, Word, PowerPoint, Tableau, Replit and copy-pasting from ChatGPT. Instead, it offers a single workspace that colleagues can access even if they haven't signed up. It has three tabs: Stage holds presentations, dashboards, simulations or mini apps; Narrative holds analyses, numbered citations and explanations; Model holds datasets and Python scripts.

Patel focused on reinforcement learning and on assessing whether AI output can be trusted, and that shapes the product's emphasis on verifiability. Users can view the Python scripts used to build documents and ask the AI chat to explain them. The system keeps memory across the individual user, the document, the team and the company. Patel says the platform is built and fine-tuned on a mix of open-source models, so it costs considerably less to run than tools built on frontier AI models.

Open format and competition

Documents can be exported as a single ".ndoc.html" file that opens in any browser. Patel plans to open-source the format, which was partly why he left Meta. His team has three people and no funding. OpenAI, Anthropic, Meta, SpaceXAI and Google all offer enterprise AI tools, and last month OpenAI launched Space, a shared workspace inside ChatGPT.

As of Tuesday, Ndoc is running an enterprise pilot and opening an alpha version to a limited number of customers from its wait list. Signing up is free, while customers pay for usage and tokens. A beta is expected in the coming weeks, with a full launch planned for next year.