What Does the AI Act Say About Using AI in Property Management?

The EU AI Act regulates AI systems by risk level and has been in force since August 2024, with a ban on certain practices such as social scoring taking effect in February 2025. For property management companies, this means: using AI for tenant communication or document analysis is permitted, as long as no automated social scoring takes place and GDPR requirements, particularly Article 22, are respected.

What is the EU AI Act, and why does it matter for property management?

The EU AI Act has been in force since August 2024. It adds a risk-tiered regulatory framework specifically for AI systems on top of GDPR, as described in GDPR and AI in property management: what you really need to know. Rather than imposing a blanket ban, the AI Act distinguishes between different risk classes of AI applications. Two categories are especially relevant for property managers: prohibited practices, and systems classified as high-risk or limited-risk, which come with additional due-diligence obligations.

Specifically, since February 2025 it has been illegal to use systems that score people based on social behavior in ways that cause harm, which is classic social scoring. No reputable provider of property management software would offer such a system. Still, when evaluating a new tool, it's worth understanding exactly how it works before signing any contract.

How do the AI Act and GDPR work together when using AI in property management?

The AI Act doesn't replace GDPR. It complements it. While GDPR governs how personal data may be processed, the AI Act classifies the AI systems themselves according to their risk potential. In practice, this means both frameworks must be considered at the same time whenever an AI system processes data belonging to tenants, owners, or applicants.

The key provision on the GDPR side is Article 22: the right not to be subject to a decision based solely on automated processing that produces legal effects or similarly significantly affects the person concerned. In practical terms, this means AI can aggregate creditworthiness information and generate recommendations, but it may not make legally binding decisions about tenancies on its own. Anyone who uses AI only for an initial assessment and keeps the final, documented decision with a human is on safe legal ground.

Other core GDPR requirements remain in place regardless of the AI Act. A data processing agreement under Article 28 is mandatory whenever an AI provider processes personal data on your behalf. Every processing activity also needs a legal basis, such as contract performance or legitimate interest. And for processing likely to pose a high risk to individuals' rights, a data protection impact assessment is required, and that is often the case for AI systems handling sensitive tenant data.

Which AI use cases are low-risk for property managers, and which are risky?

Not every AI use case in property management carries the same level of risk. Generating templates for letters and meeting minutes, as well as supporting tenant communication, are considered relatively unproblematic, as outlined in Scaling property management without adding headcount: the AI roadmap, as long as the output is reviewed by a human. Meeting moderation, legal assessments, and owner disputes, on the other hand, should stay in human hands.

Use case · Risk level · Key requirement

Phone agent for tenant calls · Low to medium · Transparency notice, data processing agreement with provider

Classifying damage reports · Low · Data minimization, no automated decision notice

Creditworthiness checks for prospective tenants · High · No fully automated final decision under Article 22

Contract analysis and clause review · Medium · Not a substitute for legal advice

Social scoring of tenants or applicants · Prohibited · Banned across the EU since February 2025

Complex legal assessments also fall into the higher-risk category: AI can summarize contract text and flag missing clauses, but it cannot replace qualified legal advice and should never have the final say on questions of tenancy law, condominium (WEG) law, or notice periods.

How are property managers using AI in practice today, in a compliant way?

AI doesn't replace people in property management. It replaces tasks, particularly the repetitive, time-consuming, error-prone routine work that currently absorbs most of the available capacity, as described in Closing the skills gap in property management: how AI helps. AI phone agents can handle a significant share of incoming tenant calls autonomously, for example verifying master data, identifying the reason for the call, and classifying damage reports by urgency.

Processing data within the EU is an important factor for a legally sound rollout. The key practical reference is the German data protection authorities' (DSK) guidance paper "AI and Data Protection" from May 2024, which provides concrete guardrails for real-world use. Companies that proceed step by step, starting with tenant communication, then documentation, then document search, can check measurable results after each phase instead of launching an uncontrolled, all-at-once project.

For property management companies that want to take this structured, GDPR-compliant approach, it's worth taking a look at a specialized solution for AI in property management, built specifically around these requirements. Those who want to dig deeper into the topic can also find further practical resources in the Academy on AI adoption and compliance.

Conclusion: what does the AI Act actually mean for your property management business?

The AI Act is not a reason to avoid AI. It's an additional framework alongside GDPR. Companies that avoid prohibited practices like social scoring, don't make fully automated individual decisions under Article 22, and meet standard GDPR obligations such as data processing agreements and data minimization can already use AI today, on solid legal footing, for tenant communication, document search, and documentation. A sober, step-by-step start in the areas that currently consume the most time will get you further than trying to build the perfect AI-driven property management operation overnight.