Govern · Governance Health Report

Is a person in charge of your AI, and can you show it?

One page answers the question your management will ask. It shows how your AI is governed and where people stepped in, for any period you choose, and downloads as a PDF to share or print.

Sign up free ⇢

Governance Health Report for the last 30 days: two things need attention, governance mode Proactive, two people who can review, one active watch word and 23 live agents, two of them on a website
The report as it opens in Analytics. Stage data.

Six answers on one page

Everything management asks about, in the order they ask it.

Setup
How governance is set up today: the mode, who can review, active watch words, live agents.
Activity
How many answers the AI gave, and where: the workspace, your website, Slack.
Human oversight
How many answers a person reviewed, and how quickly.
Corrections
What people corrected or taught, how many answers used company memory, and documents fixed or waiting.
Watch word matches
Which answers matched a watch word, and whether they were held or flagged.
Audit trail
Whether the record is intact, checked across the whole chain.

How it reports

What needs attention, never what people said.

Attention, not a score
It says what needs a look, such as reviews open for more than a week, in plain words.
Counts only
Never conversation content, so the report can travel safely.
Any period
The last 30 days or the period you pick, compared with the one before.
Company or team
Company admins see the whole company, group admins their own team.

It shows human oversight; the PDF says so plainly: "Not a legal assessment." The full record behind it is the audit trail, and the reviews it counts happen in human-in-the-loop review.

FAQ

Before you hand it to management

Who can open it, and on which plans?

Every plan includes it. Company admins see the report for the whole company, and group admins for their own team.

Does it include what people asked or what the AI said?

No. It uses counts only, never conversation content. The conversations themselves stay in the review queue and the audit trail, behind the same permissions as before.

Does it show that we meet the EU AI Act?

It shows human oversight: who can review, what was reviewed and how quickly, what was corrected, and that the record is intact. It is not a legal assessment, and the PDF says so. Whether that meets a particular rule is a question for your own adviser.

Can it come to us by email?

Yes. Admins can set up a weekly digest for the company or a group: recipients, day and hour, and which sections to include. Each person sees only what their own access allows, and can unsubscribe in one click.

Why no overall score?

A score invites arguing about the number. The report says what needs attention instead, such as reviews left open for more than a week, so someone can act on it.

Show that a person is in charge of your AI.

On every plan. Twenty minutes with the founder. No slides. The real product.

Sign up free ⇢

As featured on