Property technology providers are facing questions over the capabilities and true costs of artificial intelligence tools being marketed to estate agents, according to industry analysis highlighting concerns about transparency and compliance risks.
Riccardo Iannucci-Dawson, a data specialist at CRM provider Alto, has raised concerns about the gap between AI marketing claims and actual functionality, arguing that many recently announced tools represent basic generative capabilities rather than advanced artificial intelligence.
Three stages of AI implementation
According to the analysis, AI tools for estate agencies operate at three distinct levels. The first is generative AI, which produces listing descriptions, emails and enhanced photographs. Most tools launched in the past 18 months operate at this level.
The second stage involves insight-driven AI that analyses data from pipelines and transaction histories to surface patterns without prompting. The third and most advanced stage is execution, where AI autonomously triggers follow-ups, books viewings and manages compliance tasks.
Industry observers note that distinguishing between these capabilities matters for agencies assessing technology investments, particularly as market activity shows signs of weakening and operational efficiency becomes increasingly important.
Cost and compliance concerns
Several AI tools marketed to estate agents are built on licensed third-party large language models, which carry implications for total cost of ownership. Beyond initial licence fees, token costs accumulate each time the model processes a query, potentially creating material differences from headline pricing.
The analysis also highlights compliance risks associated with non-deterministic AI models, which can produce different responses to identical queries. In a regulated industry where agencies communicate with clients on behalf of their businesses, this variability presents operational challenges.
Questions of liability remain unresolved when AI-generated communications lead to complaints. Some providers have indicated that disclosure of AI-generated content is the agent’s responsibility rather than the platform’s, a position that may face scrutiny from regulators including The Property Ombudsman and the Information Commissioner’s Office.
Data depth as differentiator
The effectiveness of AI tools in estate agency depends significantly on the underlying data used for training, according to the analysis. Generic models trained on internet data lack specific knowledge of property transaction workflows and outcomes.
Alto claims its system draws on two decades of transaction and workflow data from agencies representing more than a third of UK estate agents. The company argues this historical depth enables its AI to identify patterns based on actual closed deals rather than generic predictions.
Industry professionals considering AI adoption are advised to verify several factors: whether announced features are currently operational or scheduled for future release, the total cost including token usage at expected volumes, who bears compliance and disclosure risks, and whether the AI was purpose-built for estate agency or adapted from generic applications.
The scrutiny comes as property investors adjust their strategies and agencies face pressure to demonstrate operational efficiency and regulatory compliance in an evolving market environment.
