Automated tenant referencing decisions affect access to a home — and the law treats that seriously
Tenant referencing platforms, letting agents and PropTech providers running automated affordability and credit checks are making decisions with clear legal effects on individuals, often at high volume and without in-house legal support. A structured assessment gives landlords and referencing platforms a defensible record for every automated tenant decision.
High volume, low legal capacity
Tenant referencing and PropTech platforms typically process large volumes of applications with automated affordability and credit-scoring checks, often run by smaller operators or letting agents without dedicated legal or compliance teams. This combination — high-volume automated decisions plus limited specialist capacity — is exactly the profile the ICO's ADM framework was designed to capture, and exactly the profile least likely to have addressed it.
An automated referencing decision that rejects an applicant based solely on a credit or affordability score, with no meaningful human review, is a strong candidate for Article 22A classification. Where the underlying data includes benefits status, disability-related income, or immigration status inferred from other checks, the special-category and fairness considerations under Article 22B and the Equality Act become directly relevant.
A referencing platform serving 200 letting agents
A referencing platform runs an automated pass/fail affordability check used directly by letting agents to accept or reject applicants, with no review step built into the product. Agents treat a "fail" result as a rejection without further consideration. Illustrative example — not a real client or verified statistic.
A single automated score determines tenancy outcomes for every applicant, with no review route.
The referencing decision is classified as ADM, safeguards are built into the product, and agents receive a documented risk assessment.
Where to start
- Classify the core pass/fail referencing decision against Article 22A before assuming it is "just a credit check."
- Build a genuine human-intervention route into the referencing product, not just a customer-service contact form.
- Document the lawful basis for processing and rule out recognised legitimate interests.
- Assess disparate-impact risk where scoring may disadvantage applicants on benefits or with non-standard income.