Blog · Published · Updated
AI usage register: how the workflow works
Short answer
An AI system register is a list of the systems in use, with an owner, the company's role, the risk status and the missing documentation. The register's rules run before the model, and a person approves the risk classification. The transparency obligations in Article 50 of the AI Act apply from 2 August 2026. At Syntalith the application costs from €3,000 net. At 40 register entries a month of 90 minutes each, it pays back in 2.0 months.
How we workDetails ↓
- 01Call30 minutes with an engineer, free
- 02PrototypeA working system on your process and data, free
- 03ImplementationTarget and fixed price in the contract; 50% at signing
- 04Acceptance50% once the agreed target is met
- analysis with an engineer and a free prototype
- €0
- team in ChatGPT or Claude, up to 10 people
- from €1,150
- automation, agent or application for one process
- from €3,000
We belong to the OpenAI and Anthropic partner programmes and learn through their official courses. What this means ↓
Author: Syntalith · 3 min read
Statistics
The AI Act and AI registers: statistics 2026
2.0 mo.
payback of a €3,000 application at 40 register entries a month: €2,094 of work today, €630 with checking and the model. Scenario.
2 August 2026
According to the European Commission (2026), the transparency obligations in Article 50 of the AI Act apply and are enforced from 2 August 2026.
2 December 2027
According to the European Commission (2026), the rules for high-risk systems in Annex III of the AI Act apply from 2 December 2027.
6
According to the Syntalith demonstrator (2026), the AI system register sorts owners, evidence and gaps for 6 supported articles of the AI Act.
49%
According to the European Commission (2026), 49% of EU SMEs and mid-caps name legal uncertainty as a barrier to AI.
Source: European Commission, Flash Eurobarometer 578 (2026) ↗
Figures checked: 22 October 2026
Who is responsible for this system and its documentation
A company needs a list of the AI systems in use, their owners and the missing material, so it can assign documentation work and see the open questions. You start with a real inventory: tools bought by departments, AI features in programs the company already has, and systems built to order. The scope of the obligations and the reading of the law need a review by the right specialists.
In our demonstrator, the register sorts systems, owners, evidence and gaps for six supported articles of the AI Act. Only requirements from the approved scope enter the register. The model prepares a draft justification and a list of gaps, but it cannot change the gate's rules. The process owner approves the risk status, the classification and the final entry.
- 01Inventorysystem, vendor, department and owner
- 02Roleprovider or deployer, approved by a person
- 03Rulesrequirements from the approved scope, applied before the model
- 04Gapsa draft justification and a list of what is missing, from the model
- 05Decisionthe owner approves the risk status and the entry
| Date | What applies |
|---|---|
| 2 August 2026 | the transparency obligations in Article 50 of the AI Act |
| 2 December 2026 | marking of content under Article 50(2) for systems placed on the market earlier |
| 2 December 2027 | high-risk systems in Annex III |
| 2 August 2028 | high-risk systems built into regulated products |
What it costs and when it pays back
- Today
- 40 register entries × 90 min = 60 h (€2,094)
- With the system, a person checks the output in 30% of the time
- €628
- Model, 2 calls per case, Claude Sonnet 5.5
- €2
- Cost of one case
- €52.35 today, €15.75 with the system
- Saving
- €1,464 a month
Payback of €3,000
2.0 months
The volume, the time and the checking share are the scenario's assumptions. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat. Model: 8,000 input tokens and 1,000 output tokens per call at Anthropic's rates, at the ECB rate of 1.1205 USD per euro on 22 October 2026.
The count covers the work around a case: searching, reading, writing and handing over. What remains is the time to check the result, 30% of today's work in the scenario. Before the quote, we measure both times on 20 to 100 cases the team has already handled, in a free test that takes 5 working days.
According to the European Commission (2026), 49% of EU SMEs and mid-caps name legal uncertainty as a barrier to AI. A register with owners comes before a policy: without an inventory, nobody knows what the policy has to cover. At Syntalith an AI audit for the AI Act and GDPR costs from €700 net and ends with a register of AI tools, a list of changes to settings and a training plan for the team. In the demonstrator, the recorded control run cost $0.087602.
What a register will not do
A register does not certify compliance and does not replace legal analysis. The six articles in the demonstrator do not represent the whole AI Act, and the legal position has to be refreshed after acts and guidelines change. A register organises the work: who is responsible for a system, what evidence already exists and what is missing. You test a register like any evaluation of a system with a model: OpenAI's evaluation best practices recommend task sets from real use and an evaluation at every change, including a change to the scope of the rules.
A few AI tools in the company
a spreadsheet with owners and dates
audit from €700
Many systems and departments, evidence in different places
a register application with rules before the model
from €3,000
Systems built to order, agents in processes
a register linked to checks on agents' actions
priced after the analysis
Where to start
The first step is an inventory of systems with their owners. The scope of the register and of the checks on agents is described in the AI Act audit for companies. We build this kind of rollout as part of AI applications. For the free analysis, a description of one case from recent weeks and of the tools it passed through is enough. You do not need to send confidential documents, and you get a written plan within 2 working days.
Worked example · Scenario
How we count the payback: an AI register in a company
We run the same count for your systems in the free analysis.
€4,188
one-off, to review the AI tools yourself: 120 h of work by IT, the data protection officer and a manager
Difference: up to €3,488
The hours and how they split for a company of this size are our assumptions. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat, for companies with 10 or more employees. The audit price grows with the number of tools, accounts and departments; for larger companies we quote it after the free review.
See the full calculation →
