Ask a policy question. The agent searches the web, reads the sources, and writes a report in the format you need. Every claim carries a citation and a confidence score, so you can see how well it is supported before you rely on it.
Runs several web searches in parallel and has Claude synthesize what it finds.
Add your own files to a library. The agent can cite them alongside web sources.
Each claim gets a citation confidence score that shows how well it is backed.
Export as a Congressional brief, policy memo, or risk assessment PDF.
Example claim written by the agent, with numbered citations that point to its sources.[1][2]
Also included: a daily digest that tracks new Federal Register items, and a multi-persona debate mode that lays out where different stakeholders disagree.
The core research agent works today. It is not public yet. We are preparing invite-only demo accounts for policy analysts and think tank researchers, and we will use their feedback to decide what to fix first.
Small teams that build AI products often do not have a governance specialist, and they find legal problems after the system is built. We plan to extend the same engine into a pre-development risk check. A developer describes the idea, and the tool reviews it against relevant laws and frameworks, such as the EU AI Act and the NIST AI RMF, and lists the compliance risks to fix before building.
If you work on AI policy and want to try the demo, send a short note with your name, organization, and the kind of questions you research.
hello@gentianadvisory.coOpen in mail app