The end of "copy and paste into ChatGPT"
Rohit Patel spent eight years at Meta working on reinforcement learning and evaluating the reliability of AI-generated outputs. In October 2026, he left his director position at Meta Superintelligence Labs to launch a radical alternative to what Microsoft and Google have done so far: Ndoc, a productivity suite built from scratch with AI as its core, not as an add-on.
The company, which operates under the same name as the product, is running an enterprise pilot and offering limited alpha access through a waitlist. Patel plans to roll out the beta version in the coming weeks, with a full launch targeted for 2027. With a team of just three people and no outside funding, the proposal is ambitious — and the intention is to prove people love the product first before seeking investment.
Three tabs, one experience
Ndoc combines three components in a single interface. The Stage tab contains presentations of key findings: slides, dashboards, simulations, or mini-apps. The Narrative tab carries analyses, numbered citations, generated figures, and the reasoning behind the numbers. The Model tab houses underlying datasets and Python scripts that manipulate them.
The design reflects Patel's background. During his time at Meta, he worked on reinforcement learning — a type of machine learning in which AI agents learn to make decisions through trial and error — and on evaluating AI output reliability. He brought that approach to Ndoc, letting users inspect the code behind their documents and ask the AI assistant to explain scripts and formulas.
"Even if you don't know any Python, you can ask the AI chat to explain, 'What does this script do? Which line has my formula?' And you can verify that formula," Patel told TechCrunch. "There's nothing hidden behind the curtain here."
In practice, users can upload a dataset and ask Ndoc to perform an analysis and write a narrative about the findings. They can also add document-specific context, including audience, tone, and memory. "It maintains a memory across you, that particular document, your team, and your company," he explained.
The format as a statement of principle
Patel's larger ambition is to reinvent the file itself. Documents can be exported as a single ".ndoc.html" file that any browser can open — a format he plans to open-source so other developers can build compatible applications without being locked into Ndoc the way so many have been locked into Microsoft or Google for years.
That desire to build an open format independently was partly what drove him to leave Meta. "At a big company like Meta, getting everyone to agree on an open-source format would undoubtedly require the input of hundreds of people and opinions on how the format should look," he said. "I wanted to move fast. Humans can only move at human pace, while... I can move at the pace that AI allows."
Why it matters
This is a bold bet. Trying to become the platform where corporate work happens is a huge wager for a founder with a team of three and no funding, especially when other AI labs are trying to own the same space. Ndoc is just another contender in a saturated market of attempts to reposition office work with AI.
But there is a fundamental difference. Competitors — Microsoft's Copilot, Google's Gemini — are AI layers applied to tools that already exist. They are features added to spreadsheets, word processors, and presentation apps that had already been consolidated over decades. Ndoc was born to be something different: a unified workspace where analytical thinking, visualization, and execution happen in the same place.
Patel doesn't pretend he doesn't understand the challenge. That's why he built connectors to import data from existing tools. "My bar was that I'm not going to release anything I don't love using," he said. "I have very optimized AI workflows for everything I do. If I start using this tool and I keep going back to my previous workflows, I haven't built a good enough tool. I have a graveyard of tools that didn't meet that bar."
The price of speed
Patel's approach reflects an almost gospel-like belief in Silicon Valley's 2026 playbook: cheap execution, fast planning. "Execute a suboptimal plan and worry about it later," he said. "By the time you make the best plan, the execution will have been done by someone else."
Ndoc is built and fine-tuned using a combination of open-source models. Patel told TechCrunch that developing the platform from scratch makes it considerably cheaper to operate than alternatives that rely on frontier AI models.
Usage and tokens will be charged after the free alpha phase. For Patel, moving as fast as possible matters more than having a perfect product from the start.
The question that remains
If AI is going to build the spreadsheets and slide decks that companies use to make decisions, someone should be able to verify its work. Ndoc's transparency — visible scripts, explainable formulas, open-source models — might be what separates this bet from the rest of the race to dominate corporate productivity. Or it could be just another tool that professionals try and abandon. The market will tell.
Sources: TechCrunch, Storyboard18, Indian Television
✓ Independent sources cross-checked and verified before publishing
