Guides / AI-written complaints / Casebriar vs ChatGPT

Casebriar vs ChatGPT for answering tenant complaints

Most letting agents already have a chatbot open, and for plenty of writing it’s the right tool. Answering a formal tenant complaint is a different job: every legal claim needs checking, every issue needs a decision, and the reply has to hold up if the complaint goes to your redress scheme. Here’s how the two compare.

Updated 4 min readEnglandGeneral information, not legal advice

In short

  • ChatGPT is a general writing assistant: fast, fluent and good for drafting and rewording.
  • Casebriar does one job: it checks each legal claim in a complaint against legislation.gov.uk, lists the records each issue needs, and drafts the reply from your decisions.
  • The difference that matters is checking. A chatbot can agree with a legal claim that isn’t in any legislation; Casebriar marks a claim it can’t support as not established.
  • Plenty of agents will use both: Casebriar for complaints, a chatbot for everyday writing.
On this page
  1. The short answer
  2. What ChatGPT does well
  3. Where a chatbot falls short on a formal complaint
  4. What Casebriar does differently
  5. When to use which

The short answer

ChatGPTCasebriar
What it’s forWriting and answering questions on any subjectAnswering formal tenant complaints in England
Where the law comes from~Its training data, or a web search in some versions✓The provision fetched from legislation.gov.uk and checked: the wording, England, and the date
A claim it can’t support×May agree with it, or cite a section that doesn’t say what it claims✓Marked “not established”, never repeated as fact
Who decides~Writes a reply straight away, deciding as it goes✓Waits for your decision on every issue
Records and deadlines×Not part of it✓Records per issue, working-day dates, your redress scheme’s details
Complaint data~Depends on the version and its settings✓Only your agency sees it; erased on your schedule
PriceFree and paid plans£13 a complaint, or plans from £39 a month

What ChatGPT does well

It writes quickly and clearly, it’s good at plain English, and it will turn rough notes into a polite paragraph in seconds. For everyday agency writing (a property description, a rewording of a letter you’ve already decided, a summary of a long email chain) it’s often all you need. Some versions can also search the web, which helps with recent news.

What it isn’t built to do is check a legal claim against the wording of a provision as it stood on the date of the events, or keep track of the records and decisions behind a formal complaint. Searching the web for an answer isn’t the same as reading the section a complaint relies on.

Where a chatbot falls short on a formal complaint

It can agree with the complaint’s law. A chatbot answering a complaint has the same weakness as one that wrote it. Two claims that come up often:

  • “Heating repairs must legally be done within 24 hours.” Section 11 of the Landlord and Tenant Act 1985 requires the landlord to keep the heating in repair; it doesn’t set a number of hours or days. A reply that accepts the 24 hours concedes something the Act doesn’t say.
  • “I’m entitled to three times my deposit.” Under section 214 of the Housing Act 2004 a court orders one to three times the deposit, on the tenant’s application, if the rules weren’t followed. It isn’t automatic.

It decides for you. Asked to answer a complaint, it will uphold or reject each point on the wording in front of it. Whether an issue is upheld depends on your records and your judgement.

Personal data. A complaint carries names, addresses and sometimes health details about the tenant and their household. Check your version’s terms and settings on keeping and reusing what you send, and whether that fits your obligations under UK GDPR, before pasting anything in. Business versions often have different terms from free consumer ones.

What Casebriar does differently

Casebriar uses AI too: Anthropic’s Claude, under commercial terms that don’t allow what you send to be used to train its models. The difference isn’t AI or no AI; it’s what happens around the model.

  1. The complaint, marked up. The tenant’s words exactly as sent, separated into issues and points.
  2. The law, checked. Each legal claim checked against the provision on legislation.gov.uk and given a result, such as verified, partial, needs evidence or not established, with the section linked.
  3. The acknowledgement. Your acknowledge-by and respond-by dates in working days, with the letter ready to send.
  4. The records. A checklist of what decides each issue, and the documents you add read for you to check.
  5. Your decisions. Upheld, partly upheld, not upheld, need more information or escalate, for every issue.
  6. The reply. Drafted from your decisions and records, with your redress scheme’s details, for you to edit and send from your own email.

It has limits too. It covers England only, it checks the topics most complaints raise (repairs, fitness for habitation, deposits, fees, rent increases, safety, harassment and eviction, access, landlord details and the Renters’ Rights Act) and says so when a claim is outside them, and it isn’t legal advice. See how it works, step by step.

When to use which

  • A formal complaint, especially a long or legal-sounding one: Casebriar. That’s the job it’s built for.
  • Everyday writing that doesn’t involve a tenant’s personal data or a legal position: a chatbot is quick and fine.
  • If you do answer a complaint with a chatbot, check every legal sentence against legislation.gov.uk yourself, check your version’s data terms first, and make each decision yourself before it writes anything.

For the other side of this, complaints tenants write with ChatGPT and what to check in them, see ChatGPT tenant complaints. For a quick first pass on any complaint, try the free complaint checker.

Sources

Legislation links go to legislation.gov.uk. Check the version in force for the dates in your complaint before relying on it. This page is general information about the law in England, not legal advice about a particular case.