LogoPropAIdir

Methodology

PropAIdir Methodology

How We Evaluate Real Estate AI Tools

PropAIdir separates public-source research from hands-on testing. Research profiles help readers understand what a real estate AI or proptech product claims to do, but they do not qualify a tool for a numeric score or ranking verdict. Only a hands-on assessment that passes our complete evidence gate can receive PropAIdir's five-dimension editorial scores.

This distinction lets readers see exactly what our editors verified, what they observed, and what remains untested. It also prevents vendor claims, public reviews, or commercial relationships from being presented as first-hand experience.

Evidence Levels

Each profile can have one of four evidence states:

  • Under Review or unspecified — the profile does not yet have a declared assessment type. It is not eligible for numeric scores or a ranking verdict.
  • Research — editors use public product pages, documentation, pricing, policies, and other attributable sources to describe features, workflows, and limitations. Research does not establish hands-on performance and is not scored.
  • Demo Observed — an editor observed a vendor-led or recorded demonstration but did not independently complete the representative workflow in a controlled account. Demo-observed profiles are not scored.
  • Hands-On — an editor independently used the product, completed a defined real estate workflow, retained the required evidence, and passed the publication gate. Only this state is eligible for numeric scores and a ranking verdict.

Current Hands-On Coverage

As of August 1, 2026, 0 of 116 publicly visible profiles have completed the hands-on evidence gate. Existing research profiles, Under Review entries, and legacy score records are not hands-on assessments and are not eligible for public numeric scores.

This is a dated coverage snapshot, not a promise that every listed product has been tested. We update the count when a profile completes the full gate. We would rather state that a tool remains untested than imply first-hand experience we do not have.

The Hands-On Evidence Gate

Before a profile can be marked Hands-On, its retained editorial record must include:

  1. The reviewer, test date, product version, and plan used
  2. A representative real estate workflow and its success criteria
  3. Controlled inputs, account state, and the conditions needed to reproduce the test
  4. Actual outputs, failures, limitations, and unusual behavior
  5. Evidence supporting every active score dimension
  6. Disclosure of affiliate, sponsor, free-account, credit, demo, or other material relationships
  7. An item-specific rationale for the final verdict

If any required element is missing, the profile remains Research, Demo Observed, Under Review, or unspecified and does not display numeric scores.

The 5 Hands-On Dimensions

A qualifying hands-on assessment is scored from 1 to 5 on five dimensions, in this fixed order:

  1. Data Accuracy — the accuracy and reliability of data or outputs observed in the tested workflow
  2. Ease of Use — the clarity of setup, onboarding, interface, and completion of the tested workflow
  3. Integration — how effectively the tested product connects with the real estate systems and data flows relevant to that workflow
  4. Pricing Value — the value observed relative to the tested plan, limits, implementation burden, and expected use
  5. Support Quality — the usefulness and reliability of documentation, onboarding, and support encountered during the assessment

Each dimension requires test-specific evidence. Historical research scores are not converted into this rubric, and one dimension's value is never copied into another. Scores are PropAIdir editorial assessments, not user ratings.

What Research Can Support

Public-source research can support factual descriptions of a product's stated capabilities, pricing, integrations, target users, terms, and documented limitations. It can also identify questions that a buyer should verify in a demo, trial, procurement review, or professional workflow. It cannot independently prove accuracy, usability, integration quality, value, support quality, or business outcomes.

No Sponsored Rankings

Vendors cannot pay for a better assessment, score, verdict, or organic ranking position. Affiliate or referral relationships, Sponsor or Featured status, paid submissions, free accounts, credits, and vendor-provided demos do not change assessment type or editorial outcomes. Material relationships must be disclosed where they are relevant.

Corrections & Vendor Feedback

If you are a vendor who believes a factual statement is inaccurate, or a reader who spots an error, contact us at editor@propaidir.com. We review correction requests against attributable evidence. A correction can update a fact, but it cannot purchase or negotiate a score.

Last reviewed: 2026-08-01