Skip to main content

How to Choose a Property Management Company: An AI-Assisted Vetting Process

By Avi Hacker, J.D. · 2026-11-19

If you are working out how to choose a property management company, AI will not make the choice for you, but it will make the comparison fair. Most owners pick a manager from two or three conversations, a headline fee percentage, and a referral. A better process sends every candidate the same written request, checks the public record, puts each fee proposal on the same all-in basis, and scores the evidence each firm supplies. A general-purpose model does the slow parts of that work: drafting the request, lining up the answers, running the arithmetic, and pointing out what a firm declined to answer. This guide is part of our coverage of AI property management and is written for owners hiring a third-party manager.

Key Takeaways

  • Send every candidate the same written request so the answers can be compared row by row. AI is most useful at flagging the questions a firm skipped or answered vaguely.
  • Check licensing and discipline at the state regulator before you read a single proposal. It is a pass or fail gate, not a scoring category.
  • Compare proposals on all-in annual cost, not the management fee rate. In the hypothetical below, the lowest rate is the most expensive proposal.
  • Ask for a real monthly owner report from a comparable property. It tells you more about the firm than any pitch deck.

Where This Fits in the Manager Relationship

Three of our guides cover other stages of the same relationship. Once you have a finalist and a draft contract, AI review of the property management agreement covers the fee base, termination rights, and performance clauses. After signing, AI reconciliation of a third-party manager's P&L covers checking the monthly numbers. This article sits before both: several candidates, no contract yet, and a decision about who gets the draft.

How to Choose a Property Management Company: Start With a Written Request

Verbal pitches cannot be compared. Each firm talks about what it does well and stays quiet about the rest. A short written request for proposal fixes that, and it is the first task to hand to a model. Give it a one-page property profile (type, unit or square footage count, location, age, current occupancy, debt, and the problems you want solved) and ask it to draft a request with numbered questions that each call for a specific, checkable answer.

Questions worth including:

  • Comparable experience: how many properties of this type and size the firm manages within a stated distance, and for how long.
  • Staffing: who the day-to-day manager would be, how many properties that person carries now, and who covers when they leave.
  • Fees: every fee the firm charges, in a table you supply, including leasing, renewal, setup, project management, and any markup on maintenance or affiliated vendors.
  • Performance history: monthly occupancy, collections, and unit turn times for two or three comparable properties over the last 24 months, with owner names removed if needed.
  • Reporting: a sample of the actual monthly owner report, and the day of the month it is delivered.
  • Money handling: where owner funds and security deposits are held, who can sign, and how often accounts are reconciled.
  • Technology and data: the software the firm runs, whether you get your own login, whether you receive a data export if you leave, and which tenant-facing tools it uses, such as the AI chatbots for property management we cover separately.
  • Ownership: whether the firm has changed hands recently or is in a sale process. Management books are bought and sold, as our guide to AI valuation of property management companies explains, and the team you interview may not be the team you get.

When the responses arrive, load them together and ask the model for one table: questions down the side, firms across the top, each cell holding the firm's answer in its own words. Then ask for a second list of every question a firm did not answer, answered with marketing language, or answered with a different metric than the one requested. That list is often the most informative thing you will read.

Check the License and the Record Before the Pitch

In many states, managing property for others is licensed activity, and the rules are not uniform. In California, leasing real property or collecting rents for others for compensation falls within the definition of a real estate broker under Business and Professions Code section 10131. Oregon has a separate property manager license, and the Oregon Real Estate Agency states that a licensed property manager must open and maintain at least one clients' trust account. Confirm the requirement with the regulator where your property sits.

Then look the firm up yourself. California's Department of Real Estate, for example, runs a public license lookup, and its site also publishes disciplinary actions taken against licensees. Check the firm, its designated broker, and the individual who would manage your property. A model with web search can help you find the right regulator page and summarize what a disciplinary order says, but read the record at the source. Models can confuse two licensees with similar names.

Fair housing compliance belongs in the same gate, because the manager's conduct can become the owner's problem. HUD's regulation at 24 CFR 100.7 states that a person is vicariously liable for a discriminatory housing practice by the person's agent or employee, regardless of whether the person knew or should have known of the conduct, consistent with agency law. Ask each candidate for its written screening criteria, its fair housing training schedule, and its complaint history, and ask your attorney how the rule applies to your ownership structure.

Credentials are useful supporting evidence. IREM has offered its Accredited Management Organization designation for management firms since 1945. Its eligibility criteria include a Certified Property Manager in an executive management position and at least three years in business, and IREM reviews each application against financial, operational, insurance, and client service standards. An accreditation tells you about a firm's processes. It does not tell you how the firm performs at your property type in your submarket.

If the property has agency debt, or will, your lender also has a view. Fannie Mae's Multifamily Guide, in Section 114.01, requires the lender to confirm that the property's management team or company manages other assets in the same geographic area and has adequate staffing and expertise managing similar multifamily assets. Ask your lender about its approval process before you narrow the list, so you do not choose a firm that then has to be approved under time pressure.

Compare Proposals on All-In Cost

Owners tend to compare the management fee percentage because it is the one number every firm states. The other fees vary far more from firm to firm. Ask the model to convert each proposal into first-year dollars using the same assumptions for your property, and to show every line.

Here is a hypothetical example with the arithmetic shown. Assume an 80-unit apartment property that collects $1,200,000 of rent per year, signs 30 new leases and 40 renewals per year, and runs $150,000 of maintenance work through the manager.

  • Firm A: 4.0% of collected rent is $48,000. A $400 fee on 30 new leases is $12,000. No renewal fee. A 10% markup on $150,000 of maintenance is $15,000. Total: $75,000, or 6.25% of collected rent.
  • Firm B: 5.0% of collected rent is $60,000. No leasing fee, no renewal fee, no markup. Total: $60,000, or 5.0% of collected rent.
  • Firm C: 3.5% of collected rent is $42,000. A $500 fee on 30 new leases is $15,000. A $150 fee on 40 renewals is $6,000. A 10% maintenance markup is $15,000. A one-time setup fee is $4,000. Total: $82,000, or about 6.8% of collected rent.

Firm C quotes the lowest rate and costs $22,000 more than Firm B in the first year. Every one of those dollars is an operating expense, so it comes straight out of NOI. These figures are illustrative assumptions, not market data.

Cost is still only one input. On the same property, a manager who collects one percentage point more of the rent brings in $12,000 a year, which is more than half of the $22,000 spread. That is why the performance history in the request matters as much as the fee table. Markups also deserve a follow-up question about who the affiliated vendors are and how work is bid, a topic covered in our guide to AI vendor management for property managers.

Score the Evidence, Not the Presentation

The sample owner report is the best single piece of evidence you will get, because it is the product you will receive every month. Sample reports usually arrive as PDFs full of tables. Anthropic's documentation on Claude PDF support explains that each page is processed as both extracted text and an image, which helps with rent rolls and ledgers laid out in columns. Ask the model to check each sample against a fixed list:

  • Income statement with budget and prior-year comparisons, and written explanations of the variances.
  • Rent roll and an aged delinquency report that agree with the income statement.
  • Bank reconciliation for each account, and a security deposit ledger that ties to the deposit account balance.
  • Check register or general ledger detail, so individual payments can be traced.
  • Notes on leasing activity, open work orders, and capital projects.

Have it report which items are present, which are missing, and where two schedules in the same package disagree. A firm whose sample does not tie out internally is unlikely to send you one that does.

Apply the same test to the performance history. Ask the model to compute average occupancy, collections as a share of billed rent, and turn times from the monthly data each firm supplied, and to flag months that are missing or figures that conflict with the narrative in the proposal. Keep in mind that this data is self-reported. AI can test whether it is complete and consistent. It cannot confirm that it is true.

References close part of that gap. Give the model the firm's proposal and ask for a call script with questions tied to its specific claims: how often the monthly report arrived late, whether the assigned manager changed, how the firm handled a disputed charge, and whether the owner would hire it again for the same property. Ask each firm for a former client as well as current ones. Then do the one check no model can do for you: visit a property the firm manages, unannounced, and call its leasing line as a prospect.

A Weighted Scorecard You Can Reuse

Set the weights before you read the proposals, so the scorecard reflects your priorities and not the best presentation. Licensing, discipline, and lender acceptability are a gate: a firm that fails does not get scored. One hypothetical weighting for the firms that pass:

  • Performance evidence at comparable properties: 30 points
  • Reporting quality, based on the sample package: 20 points
  • All-in cost: 15 points
  • Team and staffing for your property: 15 points
  • Technology and data access: 10 points
  • References: 10 points

Ask the model to propose a score for each firm in each category with a one-sentence reason that cites the proposal, then change any score you disagree with. Requiring a cited reason matters, because it keeps the model from rewarding confident prose. Record the final scores and reasons. If you have partners or lenders, the file shows how the decision was made, and you can reuse the same request the next time you buy a property.

Before uploading anything, remove resident names and personal details from sample rent rolls and ledgers, and use a business AI account whose terms cover confidentiality. If you want this set up as a repeatable process with the request template, the fee model, and the scorecard prompts, The AI Consulting Network builds these workflows for owners and operators.

Frequently Asked Questions

Q: Can AI tell me which property management company is best?

A: No. It can organize the answers, run the fee math, and flag gaps and inconsistencies in what each firm sent. The judgment about fit, the reference calls, and the site visit are still yours.

Q: How many firms should I send the request to?

A: Enough to give you a real comparison and few enough that you will read every response closely. Three to five is a workable range for most single-property searches.

Q: What should I do after I pick a finalist?

A: Ask for the firm's management agreement and review it before signing, with counsel. Fee definitions and termination terms in the contract can change the cost picture that the proposal presented.

Q: Does a professional designation mean the firm is a safe choice?

A: It is one piece of evidence about how the firm is run. Verify the state license and disciplinary record separately, and weigh the firm's results at properties like yours more heavily.

Related articles