Ex-Meta director launches Ndoc to replace Microsoft Office with unified AI

A new vision for work

Rohit Patel recently left his role as a director at Meta Superintelligence Labs after spending eight years with the company. He is now using that experience to launch a new startup called Ndoc, which seeks to fundamentally change how people work in an office setting. The goal is to move away from the current standard of using separate tools for different tasks.

Ndoc is designed to replace established workplace software like Microsoft Office and Google Workspace. Instead of just adding a chatbot to a spreadsheet, Patel wants to create a unified AI-powered platform. This workspace is intended to handle document creation, data analysis, and the presentation of findings all in one place.

The startup is currently conducting an enterprise pilot and has opened an alpha release for people on its waiting list. Patel expects to launch a beta version in the coming weeks, with a full release planned for next year. This move comes as several other high-profile executives and researchers have recently departed Meta to pursue their own ventures.

How the platform works

The architecture of Ndoc is built on three core components: Stage, Narrative, and Model. Narrative is the part of the system that handles analyses, citations, figures, and the explanations behind specific results. Model serves as the storage for the underlying datasets and the Python scripts required to generate those results.

This structure allows users to see the 'how' behind the 'what' of an AI's output. Patel believes that businesses need to be able to verify AI-generated work when making important decisions. By providing access to the scripts and data, the platform ensures there is nothing hidden behind the curtain of the AI's logic.

Users can interact with the AI to understand complex technical elements even if they lack specialized skills. For example, a user could ask the chat to explain what a specific script does or identify which line contains a particular formula. This transparency is a core differentiator from tools that simply fill in cells without showing the underlying math.

Cost and open-source strategy

To keep the platform affordable, the startup uses a combination of open-source models to build and fine-tune Ndoc. This strategy is intended to make the service considerably cheaper to operate than competitors that rely solely on expensive frontier AI models. By avoiding these high-cost models, the company can offer a more sustainable price point for businesses.

Patel also intends to open-source the platform eventually to foster a broader ecosystem. This move would allow other developers to build compatible applications without being tied to a single software provider. He wants to ensure that the technology remains accessible and flexible rather than locked into a proprietary walled garden.

The final output of an Ndoc document can be exported as a single .ndoc.html file. This file is designed to open easily in a web browser, making it easy to share. This portability is part of the effort to create a seamless experience that doesn't require users to constantly switch between different applications.

Challenging the status quo

Ndoc faces heavy competition from tech giants like Google, OpenAI, and Anthropic, all of whom are developing AI productivity tools. However, Patel argues that these companies are mostly just attaching AI features to existing spreadsheets. He describes this approach as a missed opportunity, comparing it to having electricity but only using it to put lights on bulls that till soil.

Patel argues that the current trend of 'filling cells' kills the soul of the work process. He believes the real value lies in a 'tractor' approach—a complete tool designed for the job. Ndoc aims to be that comprehensive tool by unifying the workflow rather than just decorating old software with new buttons.

Early users of the platform have expressed a specific interest in having all their work available in one location. They want to avoid the friction of moving data between different programs. By focusing on this unified workspace, Patel hopes to capture the attention of companies that are tired of fragmented workflows.

The path forward

As Ndoc moves toward its full launch next year, it will need to prove that its unified approach can compete with the massive distribution of Microsoft and Google. The success of the current enterprise pilot will be a key indicator of whether businesses are ready to move away from traditional office suites. The startup's ability to maintain its cost advantage while providing high-quality analysis will be critical.

The move also highlights a growing trend of high-level AI talent leaving major tech hubs to build independent infrastructure. Patel is joining a list of notable departures from Meta, including the Chief Global Affairs Officer and the head of product for AI for Work. This suggests a shift toward independent startups focusing on specific workplace problems.

For the reader, this means a potential shift in how professional documents are produced in the near future. If Ndoc succeeds, the way data is turned into narratives might become more transparent and integrated. It signals a move toward a world where the distinction between a 'data sheet' and a 'written report' begins to disappear.