EU AI Act risk tier & documentation register

Nobody Can Tell You Which AI Tools You Are Now Accountable For

The ad platform optimises with a model, the chatbot answers EU customers unsupervised, someone in growth wired an applicant ranker into the ATS, and half the creative comes out of a generator nobody logged. High-risk obligations became fully applicable on 2 August 2026. When somebody asks which of those systems is high risk and where the documentation is, the answer cannot be a shrug. This is the register — one row per system, with the tier, the written rationale and the review date already worked out.

Buy Now — $19 One-time payment · no subscription · works offline
Four risk tiers Guided classification Written rationale per system Overdue-review alerts CSV audit record

What the register tells you at a glance

Every system gets a tier and a reason, not just a row. The dashboard shows how much of the documentation you could actually produce today, and which classifications have gone stale.

54%
Documentation held 17 systems classified · 13 touching EU audiences · as at 19 August 2026 1 applicant-ranking system is high risk with no provider documentation on file
Documentation complete
54%
Human oversight recorded
71%
Review date in future
65%
Systems ranked by risk
High riskApplicant ranking in the ATS — Annex III employment use, vendor conformity docs still outstanding
LimitedSynthetic presenter for product explainers — disclosure missing on two social variants
LimitedWebsite support chatbot — interacts directly with people, disclosure banner live
OverdueCampaign copy generator — classified in January, never revisited after the model upgrade
MinimalInternal brief summariser — logged for completeness, literacy duty still applies

Why a classification register and not a list of tool names

A list records what you use. This works out which tier each system falls in, writes down why, and tells you what is missing before someone else finds out.

Guided questionnaire, not a guess

Twenty plain questions across prohibited practices, Annex III high-risk uses and Article 50 transparency triggers. Tick what the system actually does and the tier resolves as you answer — prohibited beats high risk, high risk beats limited, and nothing is decided by vibes.

A rationale you can hand over

Each system carries a written explanation naming the specific trigger that set its tier. That sentence is the part a reviewer actually wants, and the part nobody remembers six months after the decision.

Review dates that chase you

A classification made before a model upgrade is a classification that is wrong now. Every system gets a next-review date, and the dashboard counts the ones already overdue instead of letting them sit.

Export the record, not a screenshot

CSV for the compliance file or a shared drive, JSON for a backup you can reload later. The export carries the tier, the rationale, the oversight arrangement and the documentation score beside every system.

Built for the team that actually bought the tools

Not for counsel, and not for a governance platform nobody in marketing has a login to.

In-house marketing teams

You are the deployer of most of the AI in the building, and deployer duties are yours whether or not anyone told you. Keep one register and update it as tools come and go.

Marketing and RevOps leads

You signed the contracts for the stack. Filter by provider, see which systems carry real obligations, and chase the vendor documentation you are missing rather than all of it at once.

Agencies and consultancies

Log each client's systems separately and hand over a dated CSV with the engagement. A classification record attached to delivery is cheap to produce and awkward to be without.

Anyone preparing for a review

Client audit, board question or a regulator's letter — the useful artefact is a dated register showing what you classified, on what basis and when. That is the whole output.

“We know we use a lot of AI. We could not tell you which of it is regulated.”

Enterprise AI-governance platforms are priced for enterprises and sold to a risk team. Counsel bills by the hour and starts by asking you to list your AI systems — the list you do not have. Between those two there was nothing structured, cheap and owned by the marketing team itself — that is what this is.

One self-contained HTML file. No accounts, no API keys, no network calls, nothing uploaded. It installs to your phone or desktop home screen, works offline, and every entry stays in your own browser. Built by Mulkern AI Systems, who ship AI operations tooling for marketing teams.

A structured record-keeping aid, not legal advice. It helps you find the gaps and decide what to take to counsel — it does not replace them.

$19
One-time purchase · lifetime access · no subscription
  • Unlimited AI systems across brands and clients
  • Guided classification across all four risk tiers
  • Prohibited, Annex III and Article 50 triggers checked separately
  • Written classification rationale generated per system
  • Six-point documentation checklist scored per system
  • Provider, internal owner and data types recorded on every row
  • Human oversight measures captured in the record
  • Review and next-audit dates with overdue alerts
  • Readiness dashboard across the whole register
  • Search, filter by tier, and risk-first sorting
  • Export as CSV or JSON backup, and re-import
  • Installable PWA — works fully offline
  • Light and dark themes, mobile-ready
Buy Now — $19

Instant access after checkout. No refunds. All sales are final.