Automating rental application and leasing document checks without losing the human call

An application comes in on a Friday afternoon: the form, a photo of a driver's license, two pay stubs and a promise that the rest is coming. By Monday the leasing manager cannot remember which document was still missing, the applicant has resubmitted the wrong pay period, and the unit that should have been decided over the weekend is still open. At a property management company running anywhere from a few hundred to several thousand units, that Friday afternoon repeats on every application. Rental housing is not a small part of the market either: the Census Bureau's 2019 to 2023 American Community Survey put the number of renter occupied housing units in the United States at 44.6 million, up from 43.3 million in the prior five year period (US Census Bureau).

What does a complete leasing file actually need?

A leasing file is not complete when every document has arrived. It is complete when every document has arrived, is current, and matches what the applicant claimed elsewhere in the file.

Missing one item is the most common reason a decision gets delayed, not the screening itself. A file with six of seven documents present is not six sevenths complete. It is not ready.

Where does the process actually stall?

The delays are consistent across markets and property sizes.

• Incomplete applications. An applicant submits the form with only one pay stub when two are required, or skips the guarantor documents entirely.
• Mismatched documents. A pay stub shows a name that does not match the application, or an ID has expired.
• Back and forth with applicants. Each missing item becomes its own email or text, often from whoever on the team happened to notice it that day.
• Several people touching one file. A leasing agent collects the file, a manager checks it, and screening runs separately, with no single record of what has already been confirmed.

Which systems carry an applicant through the process?

An application typically moves through four kinds of systems before a decision is made.

Property management platforms already automate pieces of this. AppFolio's screening tools, for example, combine credit, criminal and landlord tenant history with income, employment and identity verification in one applicant record, use fraud detection built with partners including Nova Credit, Plaid and Snappt to flag altered pay stubs or mismatched details, and generate Fair Credit Reporting Act compliant adverse action notices automatically when an application is denied or approved with conditions (AppFolio). Yardi Breeze's ScreeningWorks Pro add on returns criminal, credit and eviction results in under a minute and, at its higher tier, uses automation to validate employment, banking and pay stub information as part of income verification (Yardi Breeze). Entrata and RealPage offer comparable screening and leasing modules. What none of them do on their own is decide whether to approve the applicant. That stays with the leasing manager.

What must stay with a person?

Three things belong to a person, not a workflow, and the reason is legal as much as practical.

1. The approval or denial decision. Whatever the screening result shows, the decision to approve, conditionally approve or deny an applicant is made by a person applying the same criteria to every applicant.
2. Any judgment about an applicant. An automated check can flag a mismatch or a gap. It should not be the thing that decides what that gap means.
3. Adverse action steps. When an applicant is denied, or offered less favorable terms because of information in a screening report, the Fair Credit Reporting Act requires that the applicant be told and given the information needed to dispute it. A CFPB review of complaints and applicant interviews found that landlords who took adverse action "did not consistently inform prospective tenants of their right to dispute information in reports or provide them the information necessary to do so, as required under the Fair Credit Reporting Act (FCRA)" (CFPB). The same review covers CFPB complaints about tenant screening from January 2019 through September 2022, roughly 26,700 in total, more than 17,200 of them about incorrect information appearing on a report, and concludes that some tenant screening companies are not meeting the FCRA requirement to "follow reasonable procedures to assure maximum possible accuracy" of the information they compile.

Screening decisions also carry fair housing obligations, and this is worth being careful about rather than confident about. Tenant screening companies themselves can be treated as covered entities under the Fair Housing Act: in a joint amicus brief to the Second Circuit in Connecticut Fair Housing Center v. CoreLogic Rental Property Solutions, the Department of Justice and HUD stated that HUD's guidance "applies to a wide-range of entities covered by the Act," including "third-party screening companies" (HUD and DOJ). The practical side of this does not require a legal background: apply the same written criteria to every applicant, and have a person review any automated flag before it affects a decision. This is not legal advice, no automated process should be described as compliant on its own, and a property manager's own fair housing and legal counsel should review the actual criteria and process in use.

What can automation and AI actually do here?

The realistic scope covers everything that leads up to the decision: assembling the right checklist, reading what has been submitted, checking it against the application, and preparing a complete file. It does not cover the decision itself.

Illustrative example: Consider a 400 unit apartment community during a peak leasing month. An applicant starts a lease application online, and the system immediately shows the document checklist for that unit type: photo ID, the two most recent pay stubs or an offer letter, and one prior landlord reference. As documents are uploaded, the system reads the pay stubs and checks the name and pay period against the application, flags that one stub is from a different employer than the one stated, and drafts a specific request back to the applicant asking for a corrected document, which the leasing agent reviews before it sends. Once the file is complete, it is packaged with the screening results into one record for the leasing manager, who applies the community's written criteria and makes the approval decision. The applicant hears back in a day instead of a week, and the leasing manager spends the time reviewing decisions rather than assembling files.

Picture that as a standing process rather than a one off. The lease application, the document checklist and the screening request all draw from one applicant record instead of three separate steps run by three different people. As each document lands, it is read, checked and either marked complete or flagged, with any flag or applicant message passing through a person before it goes anywhere, the same human in the loop pattern that keeps a person accountable for anything the applicant sees. The leasing manager's queue holds complete, checked files waiting for a decision, not partial files waiting for someone to notice what is missing. A similar checklist and routing pattern shows up in property maintenance coordination, one of the other workflows named on the real estate industry page: a maintenance request is checked for the details a technician or vendor needs and routed with the property information attached, instead of relayed by hand between a tenant, a manager and a vendor.

How do you measure whether it is working?

Three measures cover most of what matters, and a short baseline before any change makes the result meaningful.

• Time from application to a decision ready file. How long it takes from submission to a complete file waiting only on the approval decision.
• Touches per file. How many separate messages or hand offs it takes to get one file complete.
• Incomplete application rate. The share of applications missing something on first submission.

How should a property management company start?

1. Pick one property or unit type with enough application volume to be worth the effort, rather than standardizing every program at once.
2. Write down the current checklist for that unit type with the leasing agents who handle it today, including the exceptions they already make by hand.
3. Record a baseline for the three measures above over a few weeks.
4. Decide, before building anything, which steps are rule based and which need a person, using a table like the one above as the starting template.
5. Write down the screening and approval criteria the way you actually want them applied consistently, since that document is what any automated check should be measured against.
6. Connect the platform you already use rather than replacing it, starting with the leasing and screening tools your team already knows.

How Kastling approaches rental application and leasing document checks

Leasing document checks sit inside our AI Integration & Automation service, alongside the other real estate workflows we look at: checking applications for missing documents and fields, and routing maintenance requests to the right manager or supplier with the property details attached. An engagement usually starts with a free discovery call about the unit types, systems and documents involved, and for this kind of work a separately scoped paid audit is common, though not required for every project.

Before anything is built, we agree with the property manager which steps run automatically and which need a person's sign off, and we connect the property management, screening and e-signature systems already in use rather than replacing them. The approval or denial decision, and any judgment about an applicant, stay with the leasing team, and we agree up front how success will be measured.