A New Vision for Workplace Productivity
After an influential eight-year tenure at Meta, where he served as a director at the Meta Superintelligence Labs, Rohit Patel has officially departed the company. His next move is the launch of Ndoc, a startup aimed at fundamentally restructuring how professional work is produced and shared. Rather than offering another layer of artificial intelligence atop existing legacy software, Patel has designed Ndoc to be an entirely AI-native replacement for the conventional office productivity suites that have dominated the industry for decades.
This departure comes during a period of significant leadership shifts at Meta, following the exits of other high-profile figures such as former Chief Global Affairs Officer Joel Kaplan and AI researcher Andrew Tulloch. Patel’s transition from a tech giant to an independent founder reflects a broader trend of experts moving toward leaner, more agile environments to implement radical new software architectures.
The Unified Architecture of Ndoc
The core philosophy behind Ndoc is the consolidation of fragmented professional workflows. Modern knowledge workers typically operate across a disjointed ecosystem of tools—Excel for data, PowerPoint for presentation, Replit for coding, and generic AI chat interfaces for drafting content. Patel argues that this approach is inherently inefficient and prone to information silos.
Ndoc addresses this by integrating three distinct, interconnected components into a single workspace:
* The Stage Tab: This environment acts as the visual front-end for presentations, housing slide decks, interactive dashboards, simulations, and miniature applications.
* The Narrative Tab: Designed to provide context, this section contains the written analysis, generated figures, and the logical reasoning behind data-driven conclusions.
* The Model Tab: This is the engine room of the platform, where users can access and modify the underlying datasets and Python scripts used to generate document content.
By keeping these components within a unified interface, Ndoc allows colleagues to collaborate on a single “truth” rather than managing multiple file versions across disparate platforms.
Prioritizing Transparency and Verification
Patel’s background in reinforcement learning at Meta significantly informs the design of Ndoc. A central pillar of his engineering focus has always been the trustworthiness of AI outputs. In an era where AI-generated content can often be opaque, Patel has integrated features that allow users to audit the machine’s logic.
Because the system is built to provide transparency, users can inspect the Python scripts that drive their spreadsheets and documents. The platform is designed to explain technical processes in accessible language, allowing a user to ask the AI to identify a specific formula or explain the logic behind a calculation. This ensures that when AI produces a document or a financial projection, the human user maintains total agency and the ability to verify the data. This design removes the “black box” nature of typical AI assistants, providing a clear audit trail for enterprise decision-making.
A Shift Toward Open-Source Interoperability
A major component of Patel’s long-term strategy is the reinvention of the “file” itself. To move away from the proprietary, locked-in nature of current office suites, he has developed a format known as .ndoc.html. This file type is designed to be fully functional within any standard web browser, ensuring that recipients can view and interact with documents without needing to be registered users of the platform.
Patel plans to open-source this format, inviting other developers to build applications that support it. He asserts that this approach is crucial for breaking the reliance on closed ecosystems like Microsoft 365 or Google Workspace. His decision to pursue this independent path was driven partly by the realization that large organizations, like his former employer, struggle to reach consensus on open standards due to the sheer number of competing stakeholders and internal bureaucracy. By working independently, Patel believes he can maintain the velocity necessary to outpace larger, slower-moving competitors.
The Challenge of Disrupting Incumbent Workflows
Patel enters a crowded market where heavyweights such as OpenAI, Google, and Microsoft are attempting to embed AI into their existing products. While these companies focus on bolting AI agents onto legacy spreadsheets, Patel views this strategy as fundamentally misguided. He likens it to attaching light bulbs to farming equipment rather than building a tractor, suggesting that small, incremental improvements are not enough to change the way organizations function.
Despite this conviction, changing established human habits remains a significant hurdle. Acknowledging this, Ndoc includes connectors that allow users to import existing data from common enterprise tools. This allows users to migrate into the Ndoc environment without losing years of historical data or interrupting their immediate productivity.
Early Adoption and Future Scaling
Ndoc is currently in the early stages of its development, currently conducting an enterprise pilot and opening an alpha version to a limited group of users via a waitlist. Patel has remained disciplined in his approach, intentionally avoiding venture capital funding until he can demonstrate significant user demand and product-market fit.
The current model is free to join, with a pricing structure based on token usage. As the platform transitions toward a beta release in the coming weeks and a full launch scheduled for next year, the company will focus on refining its user experience. For Patel, the ultimate success of Ndoc will be measured by its ability to replace the “graveyard of tools” that many professionals currently use. By prioritizing speed and functional integration over perfect planning, he aims to prove that a lean, AI-native workspace can offer a vastly superior alternative to the status quo.
Disclaimer: This content is auto-generated for informational purposes only.
Source: Read Original News
