How it works

A register you can defend. By this afternoon.

Sign in, add the AI you use, answer plain-English questions. You leave with a classified register, your obligations, and an evidence trail that builds itself.

i

Add the AI you actually use.

Pick from the tools Irish teams really run, or add your own. The vendor and description pre-fill, so you just confirm how it is used. Most teams find more AI than they expected.

Magic link, no password. No card.

Add a system2 added
Search 12 common tools, or add your own

Microsoft Copilot

Productivity

ChatGPT

General AI

HireIQ — CV screening

Recruitment

+

Website chatbot

Customer-facing

+

Claude

General AI

+

Gemini

General AI

+
ii

Answer questions, not legislation.

Five plain-English screens per system. The wizard applies the EU AI Act's risk taxonomy and returns the outcome, the obligations, and the legal basis behind each one. You never open the Act.

No legal knowledge needed.

Classify · CV screening assistantQuestion 2 / 5

Does this system rank, filter or score people for a job?

For example: sorting CVs, shortlisting candidates, or scoring applicants.

Yes — it ranks, filters or scores applicants
It assists, but a person decides
No, none of these

Every answer recorded · ruleset v1.0.0

iii

Obligations, generated and dated.

Each one arrives with its deadline and legal basis. Change a status and you leave a one-line note. That note is what turns a checklist into an audit trail.

The evidence builds itself as you work.

Obligations · generated automatically1 / 3 logged
Name a human overseer · Art. 14Due 2 Aug 26
Inform candidates AI is used · Art. 50Logged
Keep input-data records · Art. 26Due 2 Dec 27

Evidence note required

Oversight assigned to A. Byrne, Head of People.

iv

Export a record a third party will trust.

One click turns the register into a formatted PDF: classifications, rationale, evidence, and the ruleset version stamped on every page. Hand it to the customer, the insurer, or the board.

Append-only. Nobody can edit it, including us.

Register · export
Export PDFBoard pack

Ardmore Talent Partners

AI Register · Statement of Position

14 JUN 2026

7 systems · 2 high-risk

SystemClassificationOblig.

HireIQ — CV screening

Shortlist applicants

High-risk3 / 5

Candidate-match model

Rank applicants to roles

High-risk2 / 4

Website chatbot

Answer site visitors

Transparency1 / 1

Microsoft Copilot

Drafting, email, docs

Minimal

ChatGPT Team

Research, drafting

Minimal

General regulatory information, not legal advice. Each entry carries its full question-and-answer record and ruleset version.

Append-only · p.1 / 9

And it stays that way

Current without becoming your job.

We track the tools and the moving law centrally. When something changes, you get a prompt to re-check what is affected. You confirm or update. Nothing drifts silently out of date.

No access to your systems, ever.

Review prompts · Q3 2026Ardmore Talent · 7 systems
Your tooling

Microsoft Copilot, in your register, shipped autonomous agents on 2 Jul. That can change its risk tier. Re-check the classification.

ConfirmUpdate
Your deadline

Your website chatbot needs an Article 50 disclosure live by 2 Aug 2026. Confirm yours is in place.

ConfirmUpdate
Your high-risk

HireIQ (high-risk) has one human-oversight obligation open since March. Assign a named owner.

ConfirmUpdate

Built from your stack, not a generic checklist

See it with your own AI.

Join the waitlist for early access — founding partners come on first.