EU AI Act · Regulation (EU) 2024/1689

Your enterprise buyer is going to ask about the AI Act. Have the answer ready.

A readiness pack for HR techs using AI in hiring: system inventory, preliminary risk classification, gaps, and an answer matrix ready for RFPs and procurement questionnaires. Delivered in 7-10 days, with a traceable methodology citing the applicable article and annex of Regulation (EU) 2024/1689.

Data handled under GDPRTraceable, versioned methodologyNo lock-in
Answer matrix for RFPs and procurement
Report section 2
Have you classified your AI systems under Annex III?
Answer ready
Yes. Inventory of 3 systems with preliminary risk classification; the screening module falls under Annex III 4(a).
Annex III 4(a) · Art. 6.2
Are you the provider or deployer of each system?
Answer ready
Documented per system: provider of the in-house matching engine; deployer of the third-party interview copilot.
Art. 25 · Art. 3(3)–(4)
How do you inform candidates that AI is involved?
Gap · P2
Notice in the application flow + a template clause for the client. Minor gap flagged: pending on the referral channel.
Art. 50 · Transparency
Extract from the deliverable. Answers ready to paste into client questionnaires.
Regulation (EU) 2024/1689

The world’s first comprehensive AI law is already in force. Your enterprise buyers already know it.

European Union flags waving in front of an EU institutional building.

The first time you feel the AI Act, it won’t be a fine. It’ll be an RFP.

Verbatim questions that enterprise procurement and legal teams are already sending to HR tech vendors:

Have you classified your AI systems under Annex III?
Procurement questionnaire · IT & ComplianceAnnex III · 4(a)
Are you the provider or deployer of each system? Where is it documented?
Vendor due diligence · LegalArt. 25 · Value chain
How do you inform candidates that AI is involved?
RFP · Responsible-AI sectionArt. 50 · Transparency
Regulation application timeline
Aug 2024
The Regulation enters into force.
Feb 2025
Prohibited practices and AI literacy already apply.
Aug 2026
Transparency (Art. 50) and general application of the Regulation.
Dec 2027
High-risk in employment (Annex III): full application.

If you sell AI-powered screening, matching, or assessment software to mid-market or enterprise clients, these questions are already landing.

Eight deliverables, one goal: pass procurement.

A single report, in three work blocks. Every conclusion references the exact article, annex, and paragraph of the Regulation.

Block 1 · Inventory and classification

Know exactly which systems you have and what you are in each

We map every AI system and model that touches your product — in-house and integrated — and set your role and risk level on a regulatory basis, not on intuition.

Deliverables included
AI system inventory every system and model, mapped.
Provider/deployer classification your exact role in each system.
High-risk analysis Annex III, employment scope: what applies and what doesn’t.
System inventory · extract
Candidate matching engine
In-house model
Provider
Automated CV screening
Annex III 4(a)
High risk
Interview copilot (third-party API)
Art. 25 · Own branding
Deployer
Each system with role, risk, and regulatory reference.
Block 2 · Obligations and gaps

What you’re missing, in what order, and how much effort

We turn the obligations that apply to you into a prioritized list of gaps with estimated effort, plus a closure plan your team can run without stalling the roadmap.

Deliverables included
Art. 50 obligations transparency toward candidates and users.
Prioritized gap report what’s missing, in what order, and with what effort.
30/60/90-day roadmap a realistic closure plan for your team.
Gap report · extract
AI notice to candidates in the referral channel
Art. 50 · Effort: low
P1 · 30 days
Document provider/deployer role per contract
Art. 25 · Effort: medium
P2 · 60 days
Register of systems and model versions
High-risk preparation
P3 · 90 days
Prioritized by sales impact and regulatory date.
Block 3 · Answers and delivery

The matrix your sales team pastes into the next RFP

The flagship deliverable: every typical procurement question with a drafted answer and its regulatory reference. Plus a 90-minute call so the whole team can defend it.

Deliverables included
RFP/procurement answer matrix answers ready to paste into client questionnaires.
90-min delivery call report walkthrough, questions, and next steps.
RFP matrix · contents
Classification and risk questions
Annex III · Art. 6
Drafted
Roles and value-chain questions
Art. 25 · Art. 3
Drafted
Transparency and candidate questions
Art. 50
Drafted
In editable format, ready for your sales team.

The inventory does not come out of a questionnaire. It comes out of your code.

The pack includes the tool we use to build it: you run it yourself, on your own machine, against your own repositories. It lists the candidate AI systems with file and line. It makes no network calls and sends us nothing — what you share is your call.

trazia-scan ~/your-repo --key=TRZ-…

Included in the pack. Node ≥ 18, no dependencies, works offline. The key identifies your engagement inside the report.

What it finds
Direct dependencies on AI SDKs in your manifests.
Inference endpoints and provider environment-variable names.
Model files, and AI tools launched as an external process.
What it does not do
It does not classify risk: Annex III turns on intended purpose, not on code.
It does not see your ATS, third-party SaaS or spreadsheets.
It does not read your line contents or your variable values.
Book a 20-min call
trazia-scan.md
openai
services/matching:14
Dependency
api.openai.com
services/matching/rank.ts:82
Endpoint
OPENAI_API_KEY
services/matching/env.ts:7
Variable
scikit-learn
services/scoring:3
Dependency
ranking.onnx
services/scoring/models:1
Model file
Does it process candidate data? Does it feed a hiring decision? That part is yours to answer.

The report ends where judgement begins: a list of questions per system that no tool can answer for you. That is what the assessment is.

One scope, one fixed price.

Not a subscription: a fixed-deliverable engagement, with a traceable methodology citing Regulation (EU) 2024/1689 and closed on a call. The price scales with scope, not with time.

Founding cohort · 3 spots left
Design Partner€1,500

The full pack for the first 3 clients, in exchange for a testimonial and feedback on the process.

Request a spot
AI Act Readiness Pack
Delivered in 7-10 daysTraceable, versioned methodologyNo lock-in
from €2,500
fixed price · excl. VAT
Scope
Standard
One product line
€2,500
fixed price
Plus
AI portfolio
€3,000+
depending on scope
Enterprise
Multiple units
Custom
scoped on the call
AI system inventory
Risk classification + gap report
RFP answer matrix
Full
Extended
Extended
Roadmap 30/60/90 + disclaimer
Traceable, versioned methodology
Number of AI systems
Up to 8
8-12
Custom
Product lines
1
Multiple
Multiple
Working calls
2
2-3
Custom

All prices exclude VAT. No lock-in. Each extra system outside the pack: €250-400.

See exactly what you’d receive

Sample report for a fictional HR tech: 10 pages with the RFP matrix as section 2.

AI Act Readiness Assessment — Trazia · Readiness report

From the first call to the report in 7-10 days.

Day 0
20-min call
Scope locked and pack fit. If it’s not a fit, I’ll tell you here.
Day 1-3
Intake + interview
A 15-20 min form and one interview to map your systems.
Day 3-6
Regulatory classification
Versioned regulatory decision tree citing the applicable article and annex of Regulation (EU) 2024/1689.
Day 7-10
Final report + delivery
Full report and a 90-minute delivery call.

Frequently asked questions

What founders, CTOs, and legal teams at HR techs ask us before we start.

There’s no “AI Act certification” for this; be wary of anyone selling one. It’s a gap assessment with preliminary classification, exactly what procurement wants to see.

Are these questions already reaching you?

Book a 20-min call

No commitment. If the pack isn’t a fit, I’ll tell you on the call.