AI dashboard and AI agents

Two screens: /ai, which shows how the AI has been used across your organisation, and /ai/agents, where you start a longer analysis and read what comes back.

The AI in Aegis reads and drafts. An agent run (one AI task you describe in a sentence and leave to work in the background) can summarise where a framework stands, reason over your risk register or draft a procedure. The AI Dashboard is a read-only health check of all that activity. The AI Agents page is where runs are started, watched and read.

The AI drafts; you decide what becomes a record

An agent produces text and stops. It never writes into a policy, a risk or a control on its own. When a run finishes you read the result and choose what to do with it: copy it, save it as a record, or turn it into an action item or work items. Each is a button a person presses.

Who uses it

Both pages need at least the Contributor role. A Viewer who opens /ai or /ai/agents is sent to the "not authorised" page. A Contributor or Manager can read the dashboard, start runs, cancel a run that is still going and act on results. An Admin also sees an extra Model execution provenance section on the dashboard and can manage AI models in Settings.

Both menu entries depend on AI assistance being enabled for your organisation. If it is switched off, the dashboard and agents entries disappear from the AI group in the left menu, and the group shows only Playbooks where that module is on.

What's on this screen

The AI Dashboard has nothing to fill in. The heading AI Dashboard sits top left with a one-line description. Top right is a purple Open AI Agents button, the only action in the page header.

Below it, the AI Agents band shows seven count cards: Total runs, then Completed, Degraded, Running, Queued, Failed and Cancelled. In the captured screen these read 74, 50, 11, 0, 0, 0 and 13; your own figures will differ.

The AI Usage band follows, with four cards: Total calls, Total tokens, Succeeded and Failed. These count every AI call in the organisation, not only agent runs. Under the cards, the By operation panel lists the first eight kinds of AI task with the calls and tokens (the units of text a model reads and writes) each has used. A Show all 102 operations line folds away the rest.

Further down, out of view in this capture, are Answer feedback and Recent agent runs. Only the first heading peeks in at the bottom edge.

Reading the AI Dashboard

Check the account name at the top right. The capture was taken as a Contributor; an Admin sees one more section on this page (see below).
Open the AI group in the left menu. It expands to show AI Dashboard and AI Agents, plus Playbooks where that module is on. Select AI Dashboard to land on this page.
Read the By operation panel. Each row names an AI task, such as AI agent (streaming) or Explain metric, with its call count and token total on the right. A row like Web search shows calls only: token use is not recorded for every operation, and "0 tokens" would read as a measured zero.
Select Show all 102 operations at the foot of the panel. The list expands in place to show every remaining operation in the same format. Select the line again to fold it.

The AI Dashboard: agent-run counts, AI usage and the per-operation breakdown — /ai.

Read the agent-run cards as a health check rather than a place to act:

Card	What it tells you
`Completed`	Finished, with a result you can open and read.
`Degraded`	The agent answered but could not reach your GRC data, so the text is generic framework material, not an analysis of your records. The run detail says so in an orange banner.
`Running`	Working in the background right now.
`Queued`	Submitted but not yet picked up. A count that stays high suggests the background worker is busy or stopped.
`Failed`	Stopped on an error. A rising count is worth raising with your administrator.
`Cancelled`	Stopped by a person before it finished. Only the `Cancel run` button sets this; nothing cancels a run automatically.

Below the usage panel

Answer feedback totals the Helpful / Not helpful votes people leave on AI assistant answers: Answers rated, Helpful, Not helpful and Rated helpful as a percentage, with no message text and no names.

Recent agent runs lists the five newest runs with their type, status and time. To open a run, use the agents page.

Model execution provenance (Admin only)

An Admin sees a Model execution provenance section between the agent and usage bands, not in the Contributor capture above. It shows which provider and model served each AI call and how much of that is backed by a gateway receipt, split By evidence status and By provider. Export provenance (CSV) downloads the detail for an auditor.

What's on the AI Agents page

Select Open AI Agents on the dashboard, or AI Agents in the AI group, to reach /ai/agents, headed AI Agents.

The upper panel, Start New Agent Run, holds an Agent Type dropdown (showing General), a wide Goal / Task Description box with an example in grey, and a Start Agent button at the bottom right. In the capture the box is empty, so Start Agent is greyed out.

The lower panel, Run History, has a Refresh control at its right-hand end and a table of every run in the organisation, newest first. Its columns are Status, Prompt, Type, Iterations, Tokens and Created. The capture shows a mix of Degraded, Completed and Cancelled runs of all three types.

Before you start a run

An agent reasons over what is already in Aegis. Check the ground first.

Select the book icon in the top bar if you want this chapter open beside the page. It opens the User Guide at the section for the screen you are on.
Confirm the account at the top right. The run is logged against it, and everyone who can open this page sees it in the history.
Make sure the risks you want examined are recorded under RISKS. A Risk Assessment run can only reason about risks already in the register; your sentence points it at them, it does not supply them.
Check COMPLIANCE for the frameworks and controls a Compliance Analysis run will reason over.
Use the AI group to move between this page and the dashboard once a run is going and you want to watch the counts change.

The AI Agents page: the run form above, the run history table below — /ai/agents.

Starting a run

Choose an Agent Type. The dropdown offers Compliance Analysis, Risk Assessment and General, and opens on General. The type sets the instructions the agent works under, which steers what it reaches for first. It is not a different engine: every run uses the same read-only lookups.
Write the goal in Goal / Task Description. Be specific. "Check compliance" gives the agent almost nothing; "list the NIS2 controls we have marked as not implemented and suggest evidence for each" gives it a direction.
Select Start Agent. It stays greyed out until the goal box has text. It then shows Starting…, resets the form to General, and reloads the history from page one with your run at the top. If the request is refused, a red message appears above the button and no run is created.

Watching a run and reading the result

A run works in the background, so nothing spins while you wait. Come back and reload the list.

Select Refresh at the right-hand end of the Run History header. The icon spins and the table reloads in place.
Read the row. Iterations counts the reasoning steps taken, and Tokens reads — until there is a figure to show.
Select the row to open the Agent Run window: status badge, agent type, the full Prompt and, once finished, the Result. A failed run shows an Error section instead. A degraded run shows an orange banner explaining that the agent could not retrieve your GRC data, so the answer is generic rather than an analysis of your records.
Open Tool calls below the result, where the agent used any. Each step names the lookup it made and shows what came back.
Act on it. Under the result, Save as record stores the text as an AI insight linked back to this run. Create action item opens a short form with the description prefilled; you write the title. Create work items appears where your role can create roadmap work. Copy in the window footer puts the text on your clipboard; it shows only for a Completed or Degraded run with an answer. A run still Queued or Running shows a red Cancel run button in the footer instead.

The table shows twenty runs to a page, with Previous / Next underneath when there are more. There is no per-person filter. Before the first run the table reads No agent runs yet; if the list cannot load, it shows Failed to load agent runs with a Try again button.

The AI assist

The form takes one free-text goal, so what an agent is good for depends on what you ask: summarising a document's obligations, suggesting a risk classification as a starting point, drafting a procedure for you to correct, or explaining what a control asks for. In every case the agent produces text and stops. Turning that text into something binding is a decision a person makes with a button. The AI assists; the organisation decides.

Tips and limits

The model can state wrong things with confidence (hallucination). The risk is highest with article numbers, recent legal changes and contract wording. Check against the official source, and read AI-drafted text line by line before it goes into a policy, an audit response or a regulatory submission.
A Degraded result is not a failed one, and nothing in it is evidence about your organisation.
"No gaps found" means the agent found none with the information it had, not that your compliance position is sound.
A submitted run cannot be edited. If the type or goal was wrong, cancel it while it is still Queued or Running and start a fresh one.
The dashboard counts are organisation-wide totals with no date filter. Only the Admin provenance section has an export.

Where this connects

AI agents goes further into run types and results. The output sits next to the records it feeds: Policies and Procedures for drafted text, Risks for a suggested classification, Compliance frameworks for the gaps a Compliance Analysis run reasons over, and Action items for tasks raised from a result. Model configuration lives in Settings; role limits are in What each role can do.