AI assist
Credo AI
Employees think about AI in varied terms. Governance teams need structured, reviewable data. This difference leads to friction on both sides: unclear submissions, repeated back-and-forth, and long review cycles.
AI assist bridges that gap by transforming natural-language descriptions into structured, governance-ready context. Freeform input → AI interpretation → Human approval.
The workflow begins by prompting the user to explain their AI system without worrying about terminology.
We then buy some time with a loading screen as the system decides what to do with the context provided.
AI-generated selections include a visible explanation. This serves two purposes: users understand exactly why a decision was made, and they can catch — and correct — any misinterpretation before it spreads downstream.
After use case creation, AI moves from an embedded layer to its parallel home accessible from the top right toolbar. The user now has a centralized place to view all AI system activity and can toggle between collaborating with AI (assist) and humans (comments).
From the AI assist panel - users can review AI activity (with referenced content in the center column) or act on AI suggestions. If AI assist is not able to do certain work, it also outlines the path of least friction to move the use case to the next governance stage.
While AI Assist appears in the UI as a single feature, it’s powered by a shared reasoning layer that blends governance data with customer context.
It is still early, but we are getting positive feedback from our design partners. More to come after we get more data and expand this feature.