Editorial Standards and Methodology
The Multifamily AI Brief documents how AI engines answer renter questions about apartment communities, where their information comes from, how their answers differ, and what multifamily teams can learn from those observations.
Our goal is not to declare winners from a single answer. It is to report what happened, explain what it may mean, and make the next step clearer.
Last updated September 2026How we conduct research
Research may combine automated collection with manual checks and review.
For each original study, we disclose the information needed to interpret the findings, including:
- The collection date or study period
- The AI engines included
- The question or type of question tested
- The market or geographic scope
- The sample size
- Relevant limitations
When the same question is tested across multiple engines, we keep the wording and research conditions as consistent as practical. Some studies include repeated runs because AI answers can change even when the question remains the same.
The specific engine set may vary depending on the purpose of the study and the availability of each tool.
How we choose renter questions
Questions are selected to reflect realistic apartment-search scenarios and are kept consistent within each comparison. They are research prompts, not a complete record of renter behavior.
How we interpret an answer
We treat several outcomes separately:
- NamedThe apartment community appeared in the answer.
- CreditedThe community or its owned website was identified as a source.
- RecommendedThe answer presented the community as an option that fit the renter’s request.
These are related but different outcomes. A community can be named without being recommended, or recommended without its own website being credited.
We do not treat one response as a permanent AI ranking. Results can vary by engine, model, date, location, personalization, available source material, and other conditions that are not always visible to the researcher.
Evidence and accuracy
We distinguish among 4 types of statements:
- 01Verified factsSupported by reliable sources
- 02Firsthand observationsWhat a specific test or review returned
- 03Interpretations and recommendationsOur reading of the available evidence
- 04HypothesesTentative ideas that require additional testing
Research findings are scoped to the test that produced them. A result from one question, market, property sample, or collection period is not presented as a universal conclusion about AI systems or the multifamily industry.
Properties, companies, and sources
We generally report findings in aggregate or anonymize individual apartment communities when their identity is not necessary to understand the result.
When a property, company, platform, or source is named, we describe the dated observation rather than making a broader characterization of that organization. The appearance of an ILS Aggregator, review platform, community website, or other source is reported as part of the AI engine’s behavior. It is not an endorsement or criticism of that source.
Fair Housing
AI engines may produce language about neighborhoods, safety, demographics, or renter suitability that requires careful handling.
When this occurs, the output may be documented as an attributed research observation. The publication does not adopt or repeat such language as its own description of a neighborhood or apartment community.
Limitations
AI answers and the property information supporting them can change frequently. Results may vary by engine, model, location, account status, personalization, and time.
Automated collection provides more consistent testing conditions, but it does not reproduce every part of an individual renter’s experience. Manual checks can provide additional context but may also reflect the conditions of a particular session.
Our findings describe what was recorded under the conditions stated. They do not prove how every renter will experience an AI tool or how an engine will respond in the future.
Disclosure
The Multifamily AI Brief is published by keepfound, a company building AI visibility software for multifamily.
Findings and interpretations are based on the evidence described in each article. The publication is written to be useful to multifamily professionals whether or not they use keepfound.