Get started
Menu

Field Report · AI-written complaints

AI-generated tenant complaints in UK lettings: the state of play in 2026

Letting agents say tenants are using AI to write longer, more legalistic complaints. This report brings together what has been published since late 2025, separates reported observations from measured data, and sets out how we plan to fill the gap. It’s for letting agents and property managers, and for journalists and researchers covering the sector.

Published Updated General information, not legal advice

Key findings

  • Agents and their trade body have been reporting AI-written tenant complaints since at least December 2025, but only as observations. We found no published count.
  • The published numbers are about complaints in general, not AI: the Housing Ombudsman made 7,082 determinations in 2024–25, up 30%, and didn’t attribute that to AI.
  • Academic research has documented AI-driven surges in submissions to government services, but not in UK lettings specifically.
  • We’re planning a survey of letting agents in England. This report will be updated with the results and the method.
On this page
  1. Methodology
  2. What has been reported
  3. The numbers that exist, and what they don’t show
  4. What nobody has measured yet
  5. How we plan to measure it
  6. Take part, or get the results
  7. What this means for letting agents
  8. Relevant law and guidance
  9. Related guides
  10. Sources
  11. Citing this report

Methodology

Source
Published material only: the Housing Ombudsman’s annual complaints review, trade press reporting, an academic preprint, and legislation on legislation.gov.uk.
Period covered
Sources published between September 2025 and August 2026.
Geography
UK lettings, with the law for England. Most reports concern England; one first-hand account is from Wales. The Housing Ombudsman covers England and deals mainly with social housing.
Evidence base
One regulator report, seven trade press articles, one academic preprint, and the Renters’ Rights Act 2025 with its commencement regulations. No survey, interview or complaint data of our own yet.
Limitations
Every lettings source describes what agents have seen, not counted. Several articles report the same August 2026 comments from ARLA Propertymark, so they aren’t independent. We don’t use the content of complaints put into Casebriar: that data belongs to the agencies using it.
Last updated

What has been reported

  1. The Housing Ombudsman’s annual complaints review for 2024–25 said landlords work in a changing environment that will include more use of artificial intelligence by both landlords and residents in the complaints process. (The Housing Ombudsman deals mainly with social housing.)
  2. Propertymark, the trade body for letting agents, described complaints becoming more polished and assertive and sometimes based on incorrect assumptions from AI tools. Its advice: break complaints into points, answer in plain language with dated evidence, and keep final decisions with a trained professional (Letting Agent Today).
  3. An antisocial behaviour officer in Wales wrote that AI-generated complaints had risen sharply in recent months, citing legislation precisely, and that housing teams weren’t resourced for the extra work (Inside Housing).
  4. Senior figures at ARLA Propertymark described a spike in long, complex complaints, some without merit, some built on outdated information such as a bill that never became law. One suggested The Property Ombudsman could ask complainants to say whether AI was used (Mortgage Solutions, LandlordZone, Landlord Today, Property Industry Eye, Letting Agent Today).
  5. Researchers published a study of “agentic flooding”: surges in demand on government services driven by cheap AI-generated text, with a dataset of 84 potential cases across 11 jurisdictions (Schmitz, Hammond and Chan, arXiv:2608.16603). It isn’t about lettings, but it describes the same mechanism.

Every lettings report above is based on what agents have seen, not on counts. That’s a reasonable starting point, but it isn’t data.

The numbers that exist, and what they don’t show

7,082Housing Ombudsman determinations in 2024–25, up 30% on the year before.Mostly social housing. Not attributed to AI. Source
84Potential cases of AI-driven surges in demand on government services, across 11 jurisdictions.Not specific to housing or the UK. Source
1 May 2026The date most of the Renters’ Rights Act 2025 took effect for private tenancies in England.New rules, and new material for mistakes. Source

You’ll see a figure in circulation claiming complaints to the Housing Ombudsman rose from about 2,600 in 2022 to over 7,000 in 2025 because of AI. We couldn’t find that in the research it’s usually attributed to, and the Ombudsman’s own reports don’t attribute its rising caseload to AI. We haven’t used it.

What nobody has measured yet

  • How common they are. The share of tenant complaints to letting agents that were written or rewritten with AI.
  • How accurate they are. How often the legal claims in them are supported by the legislation, partly supported, or not supported at all, and which claims go wrong most.
  • What they cost. How much longer they take to answer than complaints written without AI.
  • How they end. Whether AI-assisted complaints are upheld more or less often, internally and at the redress schemes.
  • What agents are doing about it. Templates, training, software, legal advice, or nothing yet.

How we plan to measure it

We’re preparing a survey of letting agents and property managers in England, covering the questions above. The method:

  • Who: people who handle tenant complaints at letting and property management businesses in England.
  • What: their own counts and estimates: complaint volumes, how many they believe were AI-assisted, time taken, outcomes and the legal claims that come up.
  • How: voluntary and anonymous at the business level. We’ll publish the questions, the sample size, the dates, and how we reached respondents, alongside the results, so the numbers can be judged on their method.
  • What we won’t use: the content of any complaint put into Casebriar. Complaint data belongs to the agencies using it, and Casebriar staff can’t read it.

Results will be published in this report, free to cite. Until then, it will be kept up to date as new reporting and data appear.

Take part, or get the results

If you handle tenant complaints in England and would like to take part, or you’d like the results when they’re published, email support@casebriar.com with “AI complaints research” in the subject. Journalists and researchers are welcome to get in touch too.

What this means for letting agents

None of this changes how a complaint should be answered. It does change how much of the work is checking.

  • Judge the claims, not the tone. A complaint that names statutes and sets deadlines isn’t more likely to be right. Check each legal claim against the provision it names, and against the version in force on the dates in the complaint. Our guide to complaints that quote the wrong legislation works through seven common mistakes.
  • Watch the dates. Most of the Renters’ Rights Act 2025 took effect for private tenancies on 1 May 2026, and the August 2026 reporting describes complaints built on outdated material, including a bill that never became law. Our page on complaints under the Renters’ Rights Act sets out what is in force.
  • Answer point by point, with dated evidence. That is the trade body’s advice reported in December 2025, and it is what a redress scheme will look for. How to respond to an AI-generated tenant complaint turns it into a process.
  • Keep a person on every decision. AI can help with the reading and checking, but the decision on each issue, and the reply, should come from someone who knows the tenancy and the records.
  • Keep your own count. Noting complaint volumes, time spent and outcomes now will tell you more about your own workload than any published figure, and is the kind of evidence nobody has collected yet.

Relevant law and guidance

This report is general information about tenant complaints in England, not legal advice about a particular case. Legislation links go to legislation.gov.uk: check the version in force on the dates in a complaint before relying on it.

Sources

Primary sources first, then trade reporting and research. We summarise reports in our own words; their content belongs to their publishers. If you cite a number, please cite its original source.

Citing this report

This report is free to cite. Please use:

Casebriar, “AI-generated tenant complaints in UK lettings: the state of play in 2026”, Field Report, updated 5 October 2026, https://www.casebriar.com/field-reports/ai-generated-tenant-complaints-2026

The address is permanent. If the findings change, we update the date and say what changed. More reports are listed on the Field Reports hub.

AI-written tenant complaints: the full set

Everything we’ve published on handling them, starting with the complete guide.

GuideAI-written complaints: the complete guideWhat they are, why they're arriving, and a process for answering one well.How toHow to respond to an AI-generated complaintA practical process for the complaint in front of you today.VerificationWhen a complaint quotes the wrong lawSeven legal claims AI tools get wrong, checked one by one.GuideChatGPT complaints: what to checkWhat a chatbot gets right, where it slips, and the checks that matter.Free toolFree complaint checkerPaste a complaint and see the laws it names and what to verify. Runs in your browser.TemplateResponse templatesAcknowledgement, final response and holding letter, free to copy.ExamplesTenant complaint examplesWorked examples: the claim, the law, the records and the reply.The lawComplaints under the Renters' Rights ActWhat's in force, what isn't yet, and what to investigate.Field ReportAI-generated tenant complaints in 2026What's been reported so far, what nobody has measured yet, and our plan to.SoftwareComplaint software for letting agentsWhat to look for, and how Casebriar compares with a chatbot or doing it by hand.ComparisonCasebriar vs ChatGPTAnswering a complaint with a chatbot or with Casebriar: what each does, and when to use which.