EU AI Act Compliance
GDPR was brutal.
Now, the EU AI Act can cost you€35.000.000
The Article 4 mandate is now active. I offer a 90-minute Aristotelian workshop to certify your staff's literacy and a Socratic audit to map your product's Annex III risk. Become compliant before fines hit.
EU AI Act Compliance
As of February 2025, you are technically in breach.
Article 4 of the EU AI Act mandates documented AI literacy for all staff who interact with AI systems, not just your engineers. Failing to certify your team is a risk you cannot afford to take.
By August 2026, high-risk AI systems under Annex III must pass full conformity assessment. Thousands of scale-ups will compete for a handful of Notified Body audit slots. The companies that prepared in 2025 will skip the queue entirely.
Only
3.5%
of companies feel fully prepared for the EU AI Act.
Source: Industry Compliance Survey 2025
An estimated €150,000 in annual compliance costs for companies that wait and rush, versus a structured pre-clearance today.
Compliance Stack
Three instruments.
One complete compliance structure.
SaaS checklist tools are soulless, missing the nuances of the AI Act. You need a comprehensive compliance structure that is both practical and effective.
Aristotelian workshop
A 90-minute fun, interactive, and educational session that fulfills the Article 4 literacy mandate in a single day.
What you learn:
- Current state of AI, most popular models and their use cases.
- AI in the business world.
- Future of AI (AGI, singularity vs doomerism).
- EU AI Act compliance, ethics, safety.
Socratic audit
Annex III review to map so-called 'Shadow AI' through Socratic questioning and deliver a 5-page memo.
What you get:
- Thorough interviews with engineers and executives.
- Identification of undocumented AI models in production.
- Annex III risk classification analysis.
- Clear 5-page pre-clearance compliance memo.
Philosophical knowledge
Technical, business, and legal disciplines combined. Your AI's existence is justified to regulators, investors, and boards.
What you achieve:
- Great understanding of AI and its implications.
- A legal and ethical foundation for your AI.
- Clear communication for board members and external investors.
- A premium EU market positioning.
About the Socratic Audit
Software cannot find your blind spots.
But I can.
Generic Compliance SaaS
- Detects what you declare, not what you actually use
- Cannot find 'Shadow AI' embedded in third-party APIs
- Generates identical PDFs regardless of true model intent
- Puts more work on your team
- Relies on you to know what to look for
- Requires you to understand the EU AI Act
Socratic Audit
- Systematic critical refutation of your AI stack
- Shadow AI discovery through rigorous socratic and technical questioning
- Precise mapping of intent vs. data architecture
- 5-page pre-clearance memo respected by regulators
- Individual expertise and accountability
Audit steps
Socratic questioning with technical expertise.
Socrates asked precise questions until the truth of ignorance and knowledge emerged.
I apply the same method to AI compliance. I don't rely on automated tools to find your blind spots. I use rigorous, Socratic-style questioning combined with deep technical expertise to challenge your assumptions and expose hidden risks.
Classification phase
Conducting conversations to determine if your system falls under the high-risk mandate, according to Annex III.
Testing phase
Auditing risk, training data, bias, robustness, technical documentation, and Human-in-the-loop oversight.
Certification
Producing a precise 5-page memo that demonstrates your compliance to regulators and boards.
Aristotle taught hylomorphism:
Every thing is a compound of matter and form. You cannot use AI responsibly without understanding the architecture behind the interface.
TheAristotelianworkshopfulfillsArticle4withaframeworkforunderstandingAI
Mechanism: understanding AI
We strip away the marketing hype to look at the matter of AI: the weights, the data, and the actual stochastic nature of the models your team uses daily.
Intentionality: imposing form
We establish the ethical and practical boundaries of how your staff interacts with AI systems. Meaning, purpose, or telos, ought to be imposed by the user, not the tool.
Certification: Article 4 compliance
Session is synthesized into a verifiable literacy certificate. 90 minutes to turn technical confusion into a documented legal asset for your board and investors.
Mechanism: understanding AI
We strip away the marketing hype to look at the matter of AI: the weights, the data, and the actual stochastic nature of the models your team uses daily.
Intentionality: imposing form
We establish the ethical and practical boundaries of how your staff interacts with AI systems. Meaning, purpose, or telos, ought to be imposed by the user, not the tool.
Certification: Article 4 compliance
Session is synthesized into a verifiable literacy certificate. 90 minutes to turn technical confusion into a documented legal asset for your board and investors.
Stockholm School of Economics — AGI Keynote (invited by SSE Debate Club)
Matei Cananau
AI compliance auditor
EU AI Act Technical Advisor
With my technical expertise, you will understand AI better than 90% of the field.
Most AI compliance consultants come from law or policy.
My background is in engineering and philosophy.
As an engineering Master's student, multiple start-up founder, philosophy reader, I am the perfect fit for conducting technical audits, interviewing your team, and training you in AI.
Audience
This engagement is
executive-grade.
I provide a technical audit of your AI systems, interview your team, and train you in AI. A packaged EU AI Act compliance deal.
Perfect For
- VC-backed scale-ups with AI inside the core product loop
- Enterprises using AI within high-risk categories
- CTOs & Heads of AI needing strictly technical compliance architecture
- Legal teams prepping engineering for 2026 conformity assessments
- Boards requiring evidence of AI governance to investors
Not A Fit
- Pre-revenue startups running LLM wrappers
- Hobbyist projects with zero commercial or PII exposure
- Companies merely looking for a generic policy PDF to check a box
Pricing
Let us talk AI.
Clear pricing, real knowledge, no hidden fees.
Team workshop
Article 4: AI literacy
Best for: passing the literacy mandate in one afternoon.
- ✦ 90-minute training for your entire team
- ✦ Board-ready Article 4 literacy certificate
- ✦ Done in one day. Remote or on-site
- ✦ A lot of fun, a lot of learning
Technical audit
Annex III: high-risk pre-clearance
Best for: proving AI system legality
- ✦ Map your exact Annex III risk level
- ✦ Hunt down and document your Shadow AI
- ✦ 5-page compliance memo for your investors and regulators
- ✦ Step-by-step plan to pass official EU checks
Continuous alignment
Stay legal as you ship
Best for: companies constantly building new AI features
- ✦ Monthly AI governance reviews
- ✦ Quarterly risk reports
- ✦ Keeping you up to date with AI
- ✦ Priority access to services
Common Questions
Everything you need to know.
Article 4 of the EU AI Act demands documented AI literacy for anyone touching AI in your company. It went into effect in February 2025. If you cannot prove your team is trained, you are in technical breach. Fines scale up to €35,000,000 or 7% of global turnover. The Aristotelian workshop makes you compliant after a 90-minute session.
Absolutely. The law targets 'deployers'. If your product uses APIs to influence decisions in high-risk categories like HR, healthcare, or credit, you fall under Annex III. Furthermore, 'Shadow AI' creates massive blind spots, as it targets employees informally using unauthorized tools in production. You are legally exposed to architectures you don't even know exist. Therefore, your executives must truly understand AI.
SaaS checklists detect what you declare, not what you actually do. They cannot interrogate an engineering team to find undocumented models hiding in third-party integrations. The Socratic audit uses rigorous technical confrontation to map the true state of your AI stack, delivering a precise 5-page memo that regulators and investors trust.
Lawyers read policy and engineers ship code. They rarely speak the same language. My background is in machine learning engineering and philosophy. I impose form upon your stochastic models according to EU law. Your engineers will respect the audit, and your board will understand the result.
August 2026. Those who wait will rush, risking an estimated compliance costs of €150,000 in panic. Much like GDPR, you want to be safe as soon as possible