US SaaS guide

EU AI Act for US SaaS Companies

A US company can still have EU AI Act exposure when it places an AI system on the EU market, serves EU customers, supports EU deployers, or produces outputs used in the EU. Start with scope, then classify the use case.

Disclaimer: This guide is for general planning by SaaS teams. It is not legal advice and should not be used as a final cross-border compliance opinion.

Scope questions for US teams

Question	Why it matters
Do you offer the AI feature to EU customers or users?	EU market availability can create review needs even when the company is incorporated in the United States.
Do EU-based customers deploy the feature inside their workflows?	Your customer may be a deployer, while your company may still have provider or supplier duties depending on the product role.
Are AI outputs used in the EU?	Article 2 includes certain providers and deployers outside the EU when output produced by the AI system is used in the Union.
Do you control the model, intended purpose, or key product behavior?	Control over intended purpose and system design is important when assigning provider, deployer, importer, or distributor roles.
Do you sell through partners, marketplaces, or enterprise resellers?	Distribution channels can add role-mapping and documentation requests.

Practical readiness checklist

Create an inventory of AI features, including internal admin tools exposed to EU customer data.
Record intended purpose, user group, model provider, data categories, and output use for each feature.
Screen for Article 5 prohibited practices before pilots or beta launches.
Screen for Article 6 and Annex III high-risk areas when the feature affects access, eligibility, employment, education, public services, safety, or rights.
Add Article 50-style transparency notices for chatbots, AI assistants, generative media, and other user-facing AI interactions.
Prepare a short customer-facing AI system note that explains the product role, limitations, oversight options, and support channel.
Review contracts and product docs so sales claims do not expand the intended purpose beyond what the system was tested to do.

Common SaaS examples to review

Feature	Review focus
AI support chatbot	Transparency notice, escalation to human support, limitation language, and data handling.
AI hiring assistant	Employment high-risk signal, human oversight, documentation, evaluation, and vendor role mapping.
Lead scoring or fraud scoring	Whether the score affects access, eligibility, pricing, services, or treatment of individuals.
Document summarization	Transparency, accuracy warnings, review path, and whether summaries are used for consequential decisions.
Generated marketing media	AI-generated content disclosure and deepfake or synthetic media rules where relevant.

Check a SaaS feature

Official sources

Last reviewed: July 3, 2026.