Former Meta AI Director Launches Ndoc
Former Meta AI researcher Rohit Patel launches an early version of Ndoc, aiming to combine analysis, code and presentations in one AI-native workspace.
Business work is often fragmented across spreadsheets, documents, dashboards and presentations. Ndoc is experimenting with a workspace that keeps the numbers and the explanation together.
What happened
Former Meta AI director Rohit Patel launched Ndoc, an early-stage AI-native productivity environment intended as an alternative to working across separate office and analytics applications.
The product uses three connected views. Stage presents outputs such as dashboards or presentations; Narrative records the analysis and explanation; and Model retains datasets and Python code underpinning the work. Users can ask AI to analyse data and produce results while retaining a way to inspect calculations.
Patel is also developing a self-contained HTML document format intended to remain accessible outside the platform. Ndoc is in limited alpha testing and enterprise pilots, with a small team. No outside financing round was announced with the launch.
Why it matters
AI tools can produce polished slides and convincing figures while obscuring how those figures were calculated. Keeping data, code and narrative connected may make automated analytical work easier to audit and update.
The larger challenge is behaviour change. Teams already depend on familiar applications and file formats, and switching costs are high. Ndoc must demonstrate that a unified workflow offers enough value to overcome those habits.
The bigger picture
A new group of startups is attempting to rebuild office software for AI agents rather than add assistants to existing products. Their opportunity is real, but success will depend on reliability, collaboration, integrations and distribution as much as model capabilities.