copilot
AR
Source-grounded AI

ImpactOS Copilot

Conversational layer over programmes, agents, FPOs, MEL studies, evidence and reports. Every answer cites rows; flag unsupported claims for human review.

Demo prompts

By role and screen

Conversation

I'm grounded on Grameen India ImpactOS data — programmes, projects, agents, FPOs, MEL studies, evidence and reports. Try a prompt from the left, or ask your own. I'll cite the rows I used.
Responses synthesised from seeded ai_insights and ai_demo_prompts rows