Ask HN: Building a ShadowAI detection tool, need inputs from the community

Hello All,

I am building a tool for detecting shadow AI (or Embedded AI). My process involves ingesting logs and classifying them as either shadow AI or not, then returning a CSV.

I want to improve it more and am looking for some input on what else I can add to the dashboard?

What would be meaningful as a data point aside from just identifying whether this tool is Shadow/Embedded AI or not.

Would appreciate any help to make it more meaningful to the end user.

Thank you!