How do you find systems nobody has listed?

Procurement records, vendor lists and expense data, plus a sweep for AI features in software already in use. Surveys find what people think of as AI; expense data finds what the organisation actually bought.

Do you make the classification decision?

We classify with the reasoning recorded, for the engaging attorney's review and adoption. Where a classification is genuinely borderline, it is flagged as borderline rather than presented as settled.

Will the register stay current?

It is built to be maintained, and delivered with a process for adding systems at procurement. We can also refresh it periodically at a lower cost than a rebuild.

AI governance legal support

EU AI Act system inventory

Before an organisation can comply with the AI Act it has to know what it is running, and the list is always longer than anyone expects. We build the inventory, classify each system, determine the organisation's role for each, and map the obligations that follow.

Send us this work →See the full rate card

Published price: $2,000–8,000 / build
Turnaround: 5 business days
Without us: $15,000–60,000

What you receive

A full inventory of AI systems in use, each classified by risk tier under the Act.

A complete inventory of AI systems, including embedded AI features in ordinary software
Risk tier classification per system, with the reasoning recorded
Role determination per system: provider, deployer, importer or distributor
Obligations mapped per system based on tier and role
A gap analysis showing which obligations are currently unmet
A maintainable register with a process for adding new systems as they are adopted

What it costs, and what it replaces

Both figures are published ranges for the same unit of work. Ours is fixed before we start; if our process gets faster, that is our gain and your price does not move.

How engagements are structured →

Price per build · same work, both sides

Adnah Legal: $2,000–8,000 / build
Typical cost without us: $15,000–60,000 / build

Roughly 87% lower at the midpoint of each published range.

How the work runs

Discovery

We work from procurement records, vendor lists and expense data rather than a survey, because the AI features embedded in software people already use are exactly what a survey misses.

Classification

Each system is classified by risk tier with the reasoning recorded, so the classification can be defended or revisited when guidance develops.

Role determination

The organisation's role per system is determined, which is where obligations actually diverge — fine-tuning a vendor model can move an organisation from deployer to provider, and that changes everything.

Obligation mapping

Obligations are mapped per system by tier and role, gaps are listed, and the register is delivered with a process for keeping it current.

What we need from you

Software and vendor inventories, procurement records and expense data
Descriptions of internally built tools and automations
Deployment context for each system: what it decides and who it affects
Vendor documentation where available

What we check before delivery

Discovery is cross-checked against procurement and expense records, not just a stated system list
Classification reasoning is recorded per system so it is reviewable
Role determinations account for modifications and fine-tuning, not just procurement status
Embedded AI features in general software are specifically swept for rather than assumed absent

When firms send us this

Organisations starting AI Act readiness with no inventory at all

Groups needing a defensible register before a customer or regulator asks

Post-acquisition integration of two organisations' AI estates

Annual refresh of a register built a year ago

Questions about eu ai act system inventory

How do you find systems nobody has listed?
Procurement records, vendor lists and expense data, plus a sweep for AI features in software already in use. Surveys find what people think of as AI; expense data finds what the organisation actually bought.
Do you make the classification decision?
We classify with the reasoning recorded, for the engaging attorney's review and adoption. Where a classification is genuinely borderline, it is flagged as borderline rather than presented as settled.
Will the register stay current?
It is built to be maintained, and delivered with a process for adding systems at procurement. We can also refresh it periodically at a lower cost than a rebuild.

Firms who send us this usually also send

AI governance

AI impact assessments

A completed impact assessment for the specific AI system, risks and mitigations documented.

$1,500–5,000 / assessment

AI governance

Internal AI policy drafting

An internal AI use policy drafted to the organisation's actual tools and risk appetite.

$1,000–4,000 / policy

Privacy

Records of processing

A complete record of processing activities, built system by system to the required schema.

$1,500–5,000 / build

Send one and judge the output.

EU AI Act system inventory at $2,000–8,000 per build, 5 business days. No minimum, no scoping call, no onboarding cycle.

Start a matter See prices and turnaround