Most SaaS companies assume the EU AI Act is about high-risk systems and does not touch them. The transparency rules tell a different story: if your product or marketing site uses an AI chatbot or AI-generated content, you are in scope from 2 August 2026.
Where SaaS products meet the Act
Three places, usually. First, the support chatbot — almost always an AI deployment that needs an interaction disclosure. Second, AI features inside the product (AI search, summaries, recommendations) which need to be inventoried and, where they generate content, labelled. Third, the marketing site itself, if it publishes AI-generated copy or media. None of these are high-risk, but all are caught by Article 50.
Why it shows up in sales
Even before enforcement, the AI Act shows up in procurement. Enterprise buyers increasingly ask vendors to confirm AI Act readiness in security and compliance questionnaires. A clear, dated answer — “here is every AI touchpoint, here is the disclosure, here is the date” — removes friction from the sale. A vague answer slows the deal.
A practical checklist
Inventory every AI system you build or deploy. Classify each against Article 50. Add an interaction disclosure to any AI a user talks to. Label AI-generated content. Write down your reasoning for the systems you decided are out of scope. Keep it all dated and reviewable. That is a defensible compliance position you can build in days, not months.