Guides
EU AI Act for SMEs and Startups: What Actually Applies
The EU AI Act applies the same rules to a 20-person startup as a multinational, but with SME-specific support: reduced conformity fees, priority sandbox access, and lower fine caps.
Article 50 transparency obligations in effect since 2 Aug 2026
EU AI Act compliance, automated.
Your buyers are already demanding AI-Act-ready assurances and ISO 42001 evidence. Compliance gaps are blocking deals today.
Hosted in the EU
Built in Berlin
The deadline moved; the work didn't
The one-off deadline was never the real value — compliance is a moving target. The Digital Omnibus proved it by shifting the high-risk dates overnight. Aikraft's regulatory-update engine flags which of your systems are affected when rules change.
How the Digital Omnibus changed the timeline → The problem
Documentation goes stale. Obligations don't.
Most companies treat AI Act compliance as a one-time project. The regulation is ongoing — and the gap between a snapshot and a living system is where the risk lives.
A snapshot, not a system
A consultant maps your systems, writes the technical file, and invoices — €50,000 to €150,000 for one high-risk system. Then a model gets retrained or a vendor gets swapped, and the file is quietly wrong. Nobody finds out until an auditor asks, or a customer’s procurement team does.
Enterprise GRC assumes a team you don't have
ServiceNow and IBM OpenPages do handle continuous compliance — with six-figure budgets, months of implementation, and someone to run it.
The deadline moved; the work didn't
Article 50 transparency applies from 2 August 2026, and grandfathering past December 2027 turns on whether a system has been substantially modified.
The solution
Everything you need to comply with the EU AI Act
Three integrated modules covering the full compliance lifecycle — from initial classification through ongoing monitoring.
Classify
Automatically classify your AI systems under EU AI Act Annex III with our guided assessment engine.
Document
Generate complete Annex IV technical documentation in minutes with AI-assisted content generation.
Monitor
Regulatory-update monitoring keeps you current as the rules move — plus drift detection and incident logging. When dates or standards change, affected systems are flagged automatically.
Covered for your role — provider, deployer, or both
Your obligations under the EU AI Act depend on whether you build an AI system or put one to use. Aikraft covers both sides, and makes clear which is which.
If you're a provider
You build or place the AI system on the market
- Annex IV technical documentation
- Risk-management and data-governance records
- Conformity-assessment evidence and EU database registration
- Post-market monitoring plans
If you're a deployer
You put a third-party AI system into use
- Fundamental Rights Impact Assessment (FRIA)
- Human-oversight assignment and records
- Operational logging and ongoing monitoring
- Serious-incident reporting workflows
See it in action
From setup to audit-ready in minutes
How it works
From zero to compliant in four steps
-
Register your AI systems
Describe each AI system you develop or deploy — name it, tag it by domain, and categorize its primary function. Aikraft supports bulk import via CSV or direct integration with your asset register. Every system gets a permanent audit trail from day one.
-
Get classified
Our guided assessment engine walks you through the Annex III criteria in plain language — no legal expertise required. Answer 10–15 questions about your system's purpose, autonomy, data use, and impact. Aikraft returns a risk tier (Unacceptable, High, Limited, or Minimal) with full reasoning you can share with regulators.
-
Generate documentation
For high-risk systems, Aikraft's AI copilot generates a complete Annex IV technical documentation package — general description, data governance records, testing methodology, human oversight provisions, and post-market monitoring plan. What takes consultants six weeks takes Aikraft minutes. You review, edit, and export audit-ready PDFs.
-
Monitor continuously
Compliance is not a project — it's an ongoing state. Aikraft monitors your classified systems for data drift, logs incidents with regulatory-grade audit trails, and alerts you when EU AI Act guidance changes in ways that affect your documented systems. Dashboards surface risk in real time. Auditors get read-only access via secure links.
Why Aikraft
Purpose-built for mid-market EU AI Act compliance
See how Aikraft compares to the alternatives.
| Feature | Aikraft Berlin, 2026 | Enterprise GRC e.g. ServiceNow | Purpose-built AI Gov e.g. Credo AI | GRC Platform e.g. Vanta, Drata | Manual / Consultants DIY + advisory fees |
|---|---|---|---|---|---|
| Pricing | €199–999/mo | €50,000+/yr | €30,000+/yr | from ~€10,000/yr | €100,000+ project |
| Setup time | Same day | 6+ months | 3+ months | Days (if already a customer) | 4+ months |
| EU data residency | ✓ | ✓ | Not publicly listed | Partial (opt-in, Enterprise tier) | ✗ |
| Self-serve signup | ✓ | ✗ | ✗ | ✓ | ✗ |
| Aikraft is built for mid-market EU teams | ✓ | ✗ | ✗ | Partial | ✗ |
| Auto-documentation Annex IV compliant | ✓ | ✓ | ✓ | ✗ — evidence mapping, not a technical file | ✗ |
| Continuous monitoring Article 72 | ✓ | ✓ | ✓ | ✗ | ✗ |
Competitor pricing and setup times are based on publicly available information as of July 2026. Where a vendor publishes indicative pricing or a range, the cell links to the source. Results may vary.
Built on trust, built for Europe
Built in Berlin
German engineering, EU values
Hosted in the EU
Your core data stays in the EU
Security built in
Encryption, access controls, and audit trails
GDPR Native
Privacy by design
From the blog
View all posts → Latest EU AI Act insights
Guides
EU AI Act Fines: The Three-Tier Penalty Structure
EU AI Act fines run from €7.5M to €35M, or up to 7% of global turnover. Here's the full three-tier structure under Article 99, what triggers each tier, and how SMEs are treated differently.
Guides
How to Build an AI System Inventory
Every EU AI Act compliance step starts with a complete inventory of the AI systems you actually run. Here's how to build one, what to include, and the gaps that trip up mid-market teams.