Rohit Patel, a former director at Meta Superintelligence Labs, has announced his departure from the company after eight years to found his new startup, Ndoc. The new platform aims to provide an "AI-native" alternative to traditional office software suites such as Microsoft Office and Google Workspace.
Through Ndoc, Patel seeks to address the workflow fragmentation currently faced by users, who are forced to switch between separate tools like Excel, Word, and PowerPoint, in addition to repetitive copy-pasting from tools like ChatGPT. Instead, Ndoc provides a unified workspace that combines data, analysis, and presentation into a single interface.
Platform Architecture: Integrating Data, Analysis, and Presentation
Ndoc's design relies on three core components that work in harmony within a single interface:
- The Stage: Dedicated to displaying final results, such as dashboards, simulations, and presentations.
- The Narrative: Focuses on providing the analysis and explanation behind the numbers, while offering numbered citations and automatically generated charts.
- The Model: Where the underlying datasets and Python scripts that power the operations are stored.
Patel emphasizes that the platform gives users a high ability to verify the accuracy of AI outputs; any user, even if they are not a programmer, can ask the AI to explain the Python script used to create a specific document or verify a mathematical equation, which removes the ambiguity surrounding the processes occurring "behind the curtain."
Why does Ndoc represent a change in work philosophy?
The fundamental difference between Ndoc and the solutions offered by major companies lies in the technical philosophy adopted. While companies like OpenAI and Google are trying to add AI tools on top of existing office systems, Patel sees this approach as ineffective, describing it as "putting headlights on oxen plowing the field instead of inventing the tractor."
Ndoc's vision relies on two main pillars:
- AI-native: Building the work environment around AI capabilities from the start, rather than as just an additional layer on top of legacy tools.
- Breaking the monopoly through open source: Patel plans to make the platform's file format (ndoc.html) open source, so that it can be opened in any browser, preventing users from falling into the "vendor lock-in" trap imposed by Microsoft and Google systems.
Ndoc uses a mix of open-source models, making its operating cost significantly lower than platforms that rely exclusively on massive and expensive models. The company is currently conducting enterprise pilots with a limited number of clients, with plans to launch a beta version in the coming weeks and a full launch next year.