The whole instrument

Every capability, in one place.

AIComply is one system, not a suite. Classification, obligation mapping, risk management, quality system, verifiable documentation, provider–deployer handoff and AI-literacy tracking — all sharing the same engine, the same evidence ledger, the same audit trail.

Classification wizard
ART. 5 · ART. 6 · ANNEX III

A wizard that walks the law, not your gut.

A branching decision tree that mirrors the Act's exact classification logic — prohibited practices first, then Annex I sectoral overlap, then Annex III §1–§8 domains, then Article 6(3) self-exclusion, then Article 6(4) authority notification when the system stays out. Every step is anchored to its article; every answer becomes a Classification record you can re-open without restarting.

  • Prohibited-practice screening (Art. 5) before anything else.
  • Annex III domain picker with conditional flags (biometrics, workplace, GPAI).
  • Art. 6(3) exception wizard with the six criteria, individually argued.
  • Art. 6(4) authority-notification record for self-excluded systems.
  • Aria suggests the answer; you confirm and sign — every decision audited.
ProviderTriage Assistant · Classify · Step 3 of 4
Step 3 of 4 · Annex III filter
Which Annex III domain applies?
ART. 6(2)
Aria suggests · §4 Employment
"Reading 'CV-Screener for hiring decisions' — Annex III §4 covers AI used in recruitment, performance evaluation and worker monitoring."
§1BiometricsRemote ID · emotion · categorisation
§2Critical infrastructureEnergy · transport · water
§4EmploymentHiring · evaluation · termination
§5Essential servicesCredit · insurance · benefits
Smart obligation mapping
ENGINE

The right obligations, auto-mapped.

Once a system is classified, the engine reads role + risk tier + Annex III domain + behavioural flags and materialises the exact obligations that apply. A workbench, not a checklist: drag tasks between To-do, In Progress, In Review and Complete. Article-derived deadlines, role-aware filters, evidence at the item level.

  • Conditional logic for biometrics, workplace AI, GPAI, emotion recognition.
  • Role-aware: deployer, provider, importer, distributor — different journeys.
  • Tabs auto-generated from the engine; phase gates lock the next step.
  • Granular checklist auto-tick from backing data (RiskPlan, Annex IV doc, Oversight plan).
CompliancePortfolio · 12 systems · Q2 2026
To do2
Art. 9
Residual risk sign-off
12 MayJD
Art. 14
Stop-mechanism designed
20 MayMK
In progress2
Art. 11
Annex IV technical doc
5 MayAS
Art. 26(7)
Worker notification
1 JunJD
Complete2
Art. 6(3)
Self-exclusion record
MK
Art. 13
IFU drafted + sealed
AS
Risk management
ART. 9

From risk register to residual sign-off.

A working RiskPlan with item-level controls, inherent and residual scoring, and a sign-off ledger. Mitigations that close items also tick the Art. 9 checklist. Reality wins — delete every risk item and the obligation flips back to NOT_STARTED.

  • Inherent + residual scoring per item, with a heat-map view.
  • Mitigation actions tied to evidence + assignee + due date.
  • Residual-risk sign-off transitions Art. 9 to COMPLETE; the ledger is permanent.
  • Periodic review (Art. 9(8)) scheduled automatically; calendar integration.
DeployerCV-Screener · Hiring · Annex III §4
Art. 9 · Risk management
Identify, analyse, mitigate.
Annex III domain confirmed
Residual risk score recorded
Post-market monitoring plan draftedOPEN →
Art. 9(8) periodic review scheduledOPEN →
Quality management system
ART. 17

QMS that lives where the work happens.

The Art. 17 quality management system is not a Word document — it's the union of every other obligation. AIComply gives you the QMS shape (policies, responsibility allocation, change management, post-market plan, incident reporting) and wires each piece to the live evidence already attached to your obligations.

  • Policies + responsibility matrix maintained inside the platform.
  • Change management triggers re-classification when systems materially change.
  • Post-market monitoring plan and serious-incident reporting workflow.
  • QMS export: a single PDF that references every backing document by hash.
CompliancePortfolio · 12 systems · Q2 2026
To do2
Art. 9
Residual risk sign-off
12 MayJD
Art. 14
Stop-mechanism designed
20 MayMK
In progress2
Art. 11
Annex IV technical doc
5 MayAS
Art. 26(7)
Worker notification
1 JunJD
Complete2
Art. 6(3)
Self-exclusion record
MK
Art. 13
IFU drafted + sealed
AS
Verifiable documentation
ANNEX IV · ART. 47 · ART. 49

Documents you can prove.

Annex IV technical files, Fundamental Rights Impact Assessments, EU Declarations of Conformity and Annex VIII registration packs assemble themselves from your live system data. Smart forms with Aria suggestions, evidence at item level, audit trail of who filled what when. Every document is hash-sealed; every dossier carries a public verify URL.

  • Annex IV builder pulls from RiskPlan, oversight plan, training data record.
  • FRIA wizard for Annex III §5 deployers; auto-skipped where not required.
  • EU DoC editor pre-fills from the chosen Art. 43 pathway.
  • Annex VIII export: the EUDB registration sheet, ready to paste.
  • verify.aicomply.ie — a regulator or buyer can check the seal without an account.
Regulator viewTriage Assistant · Dossier · DOS-000041
Annex VIII · Art. 49 dossier
Triage Assistant — provider record.
Provider
NorthLine Health Ltd · IE
Risk tier
HIGH · Annex III §5
Pathway
Annex VI · internal control
DoC
signed 18 Apr 2026 · v1.2
● SEALEDSHA-256: 8a4f…d12cverify.aicomply.ie/d/DOS-41
Provider ↔ deployer handoff
ART. 13

Art. 13 IFU exchange, in one click.

When a provider hands a high-risk system to a deployer, the Instructions for Use must travel — system purpose, performance, known limitations, oversight measures, monitoring obligations. AIComply seals the IFU package on the provider side and lets the deployer paste-import it on theirs. Cross-org, cross-account, cryptographically sealed.

  • Provider seals the IFU package; gets a one-time sharing token.
  • Deployer pastes the token; the package imports into their system record.
  • Hash chain links the deployer's copy back to the provider's seal.
  • Audit log on both sides — fields imported, who imported, when.
Provider → DeployerArt. 13 · IFU package · sealed 24 Apr 2026
Art. 13 · Instructions for use
Provider → Deployer handoff sealed.
Provider
NorthLine Health
────→
Deployer
St. James's Hospital
System purpose + intended population
Performance metrics + known limitations
Human oversight measures (Art. 14)
Monitoring obligations (Art. 26)
SEAL · 24 APR 2026 · 14:02 UTC
AI literacy programmes
ART. 4

Six role-based programmes, tracked.

Article 4 obliges providers and deployers to ensure a sufficient level of AI literacy across staff. AIComply ships six role-tailored programmes (Engineering, Product, Compliance, Data, Sales, Leadership), tracks completion per person, and surfaces the org-level percentage your auditor will ask for first.

  • Programmes calibrated to role responsibilities, not generic e-learning.
  • Per-person completion certificates; org-level dashboard for the board.
  • Refresher cadence (annual / on-role-change) tracked and prompted.
  • Certificate hash + verify URL — same trust mechanism as Annex IV docs.
Org · 84 peopleAI Literacy · Art. 4 · Q2 2026
Art. 4 · AI literacy programmes
68% of staff certified
84 PEOPLE
Engineering
92%
Product
78%
Compliance
100%
Data
64%
Sales
41%
Leadership
23%
Where to next

See how it fits your shape.

The same instrument, different starting points — by company size, by sector, by what you're optimising for.

Small Business →Enterprise →Tech Startup →Legal →Healthcare →