Running the agent assessment
Agent Discovery counts the AI agents already running in your environment — free for every organization, no subscription needed for the headcount. Answer a short questionnaire, connect or upload what you already have, run a scan; the per-agent details, the report, and one-click governing come with a plan or free trial.
Start your assessment
Open Agent Discovery in the dashboard sidebar (/dashboard/assessment) and click Start free assessment. If your assessment was set up by an Eniyan consultant instead, the page shows Assessment in preparation until their documentation request is ready; this guide covers the self-serve flow.
The questionnaire
Six one-per-step questions tailor exactly what we ask you to connect: your identity provider, cloud platforms, network egress logging, code hosting, any AI agent frameworks you suspect are in use (the one optional question), and your primary business timezone. Move with Next → and ← Back; the last step’s button reads See what to connect → and generates your evidence checklist.
Collect your evidence
The Collect your evidence checklist lists each requested artifact with a why, tagged Needed or Optional (then Provided once done). A counter tracks your progress — more sources means a more complete headcount.
- How do I get this? expands numbered, copy-paste-ready export steps for that artifact (Hide steps collapses them).
- Upload attaches the exported file (.csv .json .ndjson .log .gz .zip .txt); Replace swaps a provided one.
- For code evidence, click Connect GitHub to install the Eniyan GitHub App on the repositories you choose — the card then shows your connected account. Where GitHub linking is unavailable, upload a repo zip.
- Uploads that fail security checks are quarantined; re-export and upload again.
Run the discovery scan
With at least one source provided, click Run discovery scan. The page shows Scanning your telemetry… while the scan runs in the background; results appear on the same page when it finishes.
Reading your results
- Headcount metrics — always free: Agents found, Confirmed, No owner, and Under governance.
- Confidence tiers — each discovered agent under Discovered agents carries a confirmed, likely, or possible chip, with its source, detected providers, and a framework guess.
- Attestable chips name the provider whose cloud records can later confirm that agent’s identity cryptographically.
- Cost estimates — a card prices what governing the discovered headcount would cost at list prices on the plans that fit your account type: Personal and Team for personal accounts, or Team, Business, and Scale for team accounts, with the fitting plan tagged Suggested — no commitment implied.
- Download report PDF — the full written report; it unlocks together with the per-agent details.
Bringing agents under governance
The metrics are free; names, sources, evidence, and one-click governance come with the free trial. Click Start your free trial (14 days on Personal, 30 days on Team; nothing is billed until the trial ends). Once unlocked, click Bring under governance on a candidate: it becomes an enrolled agent, the row flips to Under governance, and attestable candidates gain an Attest → link to the new agent’s detail page under /dashboard/agents, where you can run the cloud attestation afterwards.
Re-running the assessment
One assessment runs at a time — while yours is in progress, starting another is refused. Once it completes, a Run a new assessment card appears under the results showing how many free re-runs your plan has left (Personal includes 1, Team 5, Business 10, Scale 20). Inside the allowance the button reads Run new assessment and starts a fresh one immediately; beyond it the button reads Run new assessment ($50) and a confirmation dialog states the $50 charge — added to your next invoice — before anything runs. Without a plan, re-running opens the upgrade dialog instead. Agents you already enrolled stay under governance either way.
Related: Enroll an agent and the API-first path in the Quickstart.