Updated 20 September 2026: first published on Elite Agent in May 2025. The article has been refreshed as an AI basics guide to reviewing tenant feedback and improving the way people reach your team.
An AI assistant might answer a routine enquiry quickly. But what happens when the tenant’s problem does not fit the script, or they simply want to speak with a person?
That was the question behind an experiment I ran in July 2024. After reading the New York Times article “When Your Building Super Is an A.I. Bot”, I supplied 35 reader comments to Claude and asked it to look for themes.
The exercise was interesting because AI was helping me examine people’s reactions to AI. It also suggests a useful beginner task for a property management team: organising feedback into questions you can investigate.
Can AI tell me what tenants think?
It can help you review the material you provide. That is different from establishing what tenants generally think.
My original sample came from self-selected readers commenting on one US news story. They were not a representative survey of tenants. The themes Claude suggested included interest in convenience, concerns about privacy and a desire to reach a person when a problem became complicated.
Those were starting points to check against the comments. They did not establish how common each view was in the wider population, or prove that different generations wanted different services.
The same distinction matters with your own feedback. Five complaints about communication may reveal something worth fixing. They cannot, by themselves, tell you how every tenant feels.
Start with a small, usable sample
Choose feedback you are entitled to analyse and use a tool your agency has approved for that information. Remove identifying details before supplying it. Publicly visible comments can still contain personal information.
Keep the original wording and give each comment a simple reference such as T01 or T02. Note how the feedback was collected and when. Separate a routine survey from complaints received after a particular incident; combining them can obscure what you are trying to understand.
Decide the question first. “Where do people become unsure about the next step in a maintenance request?” gives you something more useful to investigate than “Do tenants like us?”
Try this feedback-review prompt
Help me review this authorised, de-identified property-service feedback.
Question we want to investigate: [QUESTION]
How and when feedback was collected: [CONTEXT]
Feedback with anonymous reference IDs: [COMMENTS]
Identify recurring themes, contrary views and missing context. For each theme, list the supporting comment IDs and count distinct comments, not repeated mentions. A comment may support more than one theme; explain that overlap.
Separate what people explicitly said from your interpretation. Do not infer age, demographics or motives, or generalise this sample to all tenants. Flag ambiguous comments for human review.
Suggest three questions to investigate with our team and one small service change we could test. Include how customers can reach a person. Do not claim the change will improve satisfaction without evidence.Check the summary before acting on it
Open the comments behind each theme. Does the summary preserve their meaning? Did the model miss a contrary view or count the same comment twice?
Consider a fictional example: several people say they received an acknowledgement but did not know when anyone would contact them. That points to a question about follow-up, not necessarily a demand for a new chatbot.
Ask the team what actually happens after acknowledgement. Check who owns the next step, how the tenant is told about it and what happens when that person is unavailable.
Test one improvement with real people
You might clarify the next-step message or make the route to a person easier to find. Agree who will review the change and which evidence would help you judge it, such as repeated questions about the status of a request.
Ask customers how they prefer to contact you rather than guessing from their age. Review the feedback again after the trial, keeping its collection method and limits visible.
The useful result is a better question and an accountable next step. AI can help organise what people have told you; listening and responding remain your team’s work.
The original version appeared in Samantha McLean’s AI Powered Agents newsletter. Read the refreshed source article on AI Powered Real Estate by Samantha McLean on Substack.