Skip to content

AI Act · deterministic engineYour AI compliance, with a deterministic engine.

AI prepares your inventory. You confirm the facts. Obligations follow explicit, traceable and reproducible rules.

For legal, compliance and AI teams. In companies and consulting.

( roles, obligations and documents: everything in its place )
  • Obligations from explicit rulesDeterministic engine, accessible sources
  • Organised evidenceDocuments linked to obligations
  • Accessible historyVersions and activities recorded
The deterministic engine

An LLM can make mistakes.
Obligations follow rules.

LLMs can invent information or give inaccurate answers. Aigolex uses them to assist data gathering. A deterministic engine determines obligations from the facts you confirm.

  • Explicit rules and accessible legal sources
  • Reproducible results with the same data, date and versions
  • Missing facts and cases needing review are flagged
Try the method in 5 questions
Two different roles, one process

AI proposes

Extracts information from your files and prepares the inventory.

Information to review
You check and confirm the facts

The engine determines

Applies versioned rules to the role, system and context. No LLM decides obligations.

ObligationsSources and reasoningMissing information
( verifiable rules, based on facts you confirm )
The problem

AI evolves. Governance needs to keep up.

  1. Clarify roles and responsibilities: obligations start here.( the wrong role changes the obligations you manage )
  2. Obligations depend on use. “It’s only ChatGPT” is not enough.
  3. March's perfect documentation is April's archaeology.
  4. Auditors, clients and the board need answers. “We’re working on it” is not enough.

AI Act compliance is not a PDF you file away. ( It is maintenance, forever. )

How it works

Systems and uses. Compliance starts here.

Aigolex connects your company, operational areas, systems and uses in one inventory.

  1. Company

    The legal entity responsible.

  2. Operational areas( workspace )

    Where AI is developed, tested or used.

  3. AI systems( asset )

    AI systems and models, including general-purpose AI.

  4. Uses( deployment )

    How, by whom and for what each system is used.

( the use context matters, alongside the system and your role )

Aigolex AI

From your files
to your inventory. With AI.

Start with repositories, documents and spreadsheets. AI proposes systems and uses for you to review and confirm.

From sources to assets and deployments

Your sources

GitHub</>
GitLab</>
Bitbucket</>
Google Workspace
Microsoft 365
Excel and CSVXLSX
PDF documentsPDF
Word and textDOCX

AI extracts and organises

LegisClaudeChatGPTOther LLMs

Your inventory

AssetProposal
CV screening

README.md Source

Deployment
HR Candidate selectionUse linked to the system

processes.docx Source

Review and confirm
LegisBuilt in

Import with Legis.

Connect repositories and document services or upload files. Legis extracts information, cites sources and proposes your inventory.

Less manual entry, from data you already have.
ClaudeChatGPTOther LLMs
Connected

Use the assistant you know.

Connect Claude, ChatGPT or an MCP assistant. It reads available sources and sends the proposal to Aigolex.

Use your subscription, without consuming Aigolex import tokens.
Features

From inventory to managing obligations.

Scope

Find where you use AI. Even where you didn’t expect.

  • AI prepares your inventory from code, documents and spreadsheets
  • Review and confirm systems (assets) and uses (deployments)
  • Clarify your role: provider, deployer or both
( the right role is the starting point )

Obligations

Obligations identified by rules, verifiable in the sources.

  • A deterministic engine applies rules to confirmed facts
  • Role, risk and application dates guide your checklists
  • Missing facts and cases requiring review stay explicit
( every obligation has a basis and a record )

Evidence

The dossier to have before the auditor emails.

  • Documents linked to operational areas, systems and uses
  • Versioned system descriptions and risk assessments
  • Documented human oversight and monitoring
( a folder on Drive is not a management system )

Continuity

Follow changes. Know what to review.

  • Track changes to systems, uses and operational areas
  • Review obligations and risk when facts change
  • Keep documents and assessments aligned
( compliance continues after the first assessment )
Free assessment

Five questions. See where to start.

A first direction on a few areas to explore. Around 2 minutes, instant results and no email required.

Try the quick assessment
Sample question

Do you use AI to screen or evaluate job candidates?

( a use to assess, including its conditions and exceptions )

For those who govern AI, in companies and consulting.

For companies

Systems, uses and obligations shared across business functions.

  • is based or operates in the European Union
  • uses AI systems in its processes
  • needs to document and monitor AI governance

For professionals and consultants

A common method for teams and consultants, across companies.

  • Legal and compliance
  • Risk and trust
  • AI leads and technical teams
  • Advising companies on the AI Act
Method and traceability

A clear method. Work you can verify.

Obligations follow deterministic rules. Sources, evidence and history make the process verifiable.

  • Explicit rules and sources

    The engine links obligations to legal references and declared facts.

  • Linked evidence

    Documents and evidence linked to systems, uses and obligations.

  • A history of changes

    Versions and the activity log show what changed and what needs review.

Compliance dossierIllustrative example

Candidate CV screening

HR · System use · Role: deployer
  1. Legal referenceReg. (EU) 2024/1689 · Art. 26
  2. Applicable obligationHuman oversight
  3. Linked documentHuman overseer appointmentVersion 2 · Published on 14 March
Published versions and changes available for review
Example: from a legal source to a document and its version.

( from the legal reference to the evidence, without losing the thread )

FAQ

The AI Act, without the hedging.

What is Aigolex?

Aigolex organises systems, uses, obligations and evidence for AI Act compliance. AI proposes your inventory from files; you confirm the facts and the deterministic engine identifies obligations. Documents and versions form an exportable dossier.

Who does the AI Act apply to?

It applies to those who place AI systems on the market or put them into service in the European Union (providers), to those who use them for professional activities in the Union (deployers), and to providers and deployers established outside the EU when the output of the system is used in the Union. It also covers importers, distributors and authorised representatives (Article 2).

What is the difference between a provider and a deployer?

A provider develops an AI system, or has it developed, and places it on the market under its own name or trademark. A deployer uses an AI system under its authority in the course of a professional activity (Article 3). A company using ChatGPT or HR software with AI features is, as a rule, a deployer.

If we only use ChatGPT, does the AI Act apply to us?

Yes, at least in part. Since 2 February 2025, organisations using AI systems must ensure a sufficient level of AI literacy among the staff who use them (Article 4). Other obligations depend on the use: for example, using AI to screen job candidates is one of the high-risk cases listed in Annex III.

When do the obligations apply?

The Regulation has been in force since 1 August 2024 and applies in stages (Article 113): prohibitions and AI literacy from 2 February 2025, obligations for general-purpose AI models from 2 August 2025, most other provisions from 2 August 2026, and the rules for high-risk systems embedded in regulated products (Annex I) from 2 August 2027. These are the dates in the original text: the EU legislator may amend them.

What are the penalties under the AI Act?

Up to €35 million or 7% of worldwide annual turnover for prohibited practices, up to €15 million or 3% for breaches of other obligations, and up to €7.5 million or 1% for supplying incorrect information to authorities (Article 99). For SMEs and start-ups, the lower of the two amounts applies.

Does Aigolex replace a legal advisor?

No. Aigolex organises obligations, evidence and deadlines. Legal assessments remain with professionals, including advisors supporting several companies.

Does Aigolex use AI to create my inventory?

Yes. Legis reads repositories and files, cites sources and proposes systems, uses and answers. You review and confirm before anything is created. AI assists data gathering; the deterministic engine identifies obligations.

Can I connect Claude or ChatGPT?

Yes, with an MCP-compatible assistant. Create the connection in Aigolex settings: the assistant reads sources it can access and sends a proposal for your confirmation. Use your subscription, without consuming Aigolex import tokens.

Does an LLM decide my obligations?

No. A deterministic engine applies explicit rules to confirmed facts. With the same data, assessment date and rule version, the result is reproducible. LLMs can make mistakes or invent information, so their proposals need review. The engine flags missing facts and cases requiring review, with references to legal sources.

Bring order to your AI governance.

( start with your data, reduce manual work )