After eight years working on artificial intelligence at Meta, Rohit Patel has left the company to build Ndoc, a startup developing an AI-native alternative to familiar office software. Its aim is to bring analysis, documents and presentations into one shared workspace rather than layering AI features onto separate apps.
Ndoc is organized around three areas: Stage for presentations, dashboards and simulations; Narrative for analysis and explanations; and Model for datasets and Python scripts. Users can upload data and ask the system to generate analysis and a written account of its findings, with context such as audience and tone carried across projects and teams.
Making AI work easier to inspect
A central part of Patel's approach is transparency. Ndoc exposes the scripts used to create its outputs, allowing users to inspect how a result was produced or ask the AI to explain a formula. Patel says this is important when businesses rely on AI-generated spreadsheets and presentations to support decisions.
The platform is built using a combination of open-source models. Patel also plans to offer documents as a single .ndoc.html file that can open in a browser, and to open-source the format so developers can create compatible tools.
Ndoc is currently running an enterprise pilot and offering a limited alpha to people on its waitlist. A beta is planned for the coming weeks, with a broader launch targeted for next year. The startup has a three-person team and is seeking customer feedback before pursuing funding. Connectors are intended to help organizations bring in information from tools they already use.
By combining creation, analysis and inspectable code in one place, Ndoc reflects a broader shift toward AI systems that can manage complete workflows. If its open format gains adoption, it could also encourage more flexible ways to create and share workplace documents.
