Can Your Auditor Abstract Your Leases?

Short answer: usually yes, but only under conditions that most engagement letters handle badly. Abstracting a client's leases is a nonattest service. Under the AICPA Code of Professional Conduct, a firm may perform it for an attest client only if the general requirements in ET section 1.295.040 are satisfied: the firm takes on no management responsibilities, and the client designates a competent person to oversee the work, evaluates the results, and accepts responsibility for them. Get those wrong and independence is impaired, whatever the workpapers say.

This question comes up constantly during ASC 842 work, because lease abstraction sits in an awkward spot. It is tedious document reading rather than accounting judgment, so it feels safe to hand to the auditor who is already reading the file. That instinct is right about the nature of the work and wrong about how easily the safeguards get satisfied in practice.

Why lease abstraction counts as a nonattest service

Reading a lease and summarizing its terms into a schedule is a nonattest service in the same family as bookkeeping and financial statement preparation. You are producing accounting records the client will rely on. That does not make it prohibited. The AICPA framework assumes firms perform nonattest services for attest clients all the time and sets conditions rather than a ban.

The distinction that matters is between reading and concluding. Pulling the stated commencement date, the rent exhibit, and the renewal notice window off the page is reading. Deciding that a five year renewal option is reasonably certain of exercise, or that the rate implicit in the lease is not determinable so an incremental borrowing rate applies, is judgment. The first is a service you can perform. The second is management's, and performing it for them is where independence actually breaks.

The four conditions in ET 1.295.040

Before the work starts, the client must agree to all of the following:

• Assume all management responsibilities. The firm does not make decisions on the client's behalf or take on a role that management should occupy.
• Oversee the service by designating an individual with suitable skills, knowledge, or experience. Usually shortened to SKE, and usually the condition that fails.
• Evaluate the adequacy and results of the services performed. Not acknowledge. Evaluate.
• Accept responsibility for the results. The schedule becomes the client's, not the firm's.

The understanding should be documented in writing, and the documentation should be specific about what the firm is doing. "Assistance with lease implementation" is not a description of a service; it is a description of an ambiguity.

Where firms actually trip: the SKE test

Nobody fails the first condition on purpose. The failure is almost always the second one, and it is quiet. A firm names the client's controller as the SKE individual, the controller signs the representation, and everyone moves on. Then the question is asked at peer review: could that person actually have evaluated this lease population?

If the deliverable is a spreadsheet of numbers with no link back to the underlying documents, the honest answer is often no. The controller cannot test a commencement date they have no practical way to trace, across a hundred leases, in the time available. They signed for oversight they were not equipped to exercise, and the designation was a formality.

This is the strongest practical argument for source linked extraction, and it has nothing to do with speed. When every extracted field cites the page and clause it came from, a client reviewer can genuinely spot check the population: pull twelve leases, click through to the cited page, confirm the term and the notice window in seconds each. That is evaluation someone can defend, and it produces evidence that the oversight was real.

What your auditor cannot do

Regardless of how carefully the conditions are met, some things stay off the table on an attest engagement:

• Concluding, on the client's behalf, whether renewal or termination options are reasonably certain of exercise.
• Selecting the discount rate the client will use, including making the risk free rate election available to private companies.
• Deciding whether a contract contains a lease at all, which is the judgment that drives the entire embedded lease population.
• Determining lease classification, or authorizing and posting the resulting entries.
• Serving, in substance, as the client's accounting department for lease matters on an ongoing basis.

A firm can absolutely explain the guidance, lay out the alternatives, and tell a client what peers typically conclude in similar fact patterns. What it cannot do is make the call and then audit it.

Does using AI change the independence analysis?

Not in principle. Independence rules care about who performed the service and who took responsibility for it, not which tools were used. If the firm runs the extraction, the firm performed a nonattest service and the same four conditions apply. Automation does not convert a service into a non service.

What automation changes is the evidence. A tool that returns fields with page citations makes the client's oversight testable in a way a manually typed spreadsheet never was, and it shifts effort from producing the data to reviewing it. It also sharpens the reading versus concluding line, because a well built extraction returns what the document says and declines to guess at what management intends. When the tool refuses to conclude that an option is reasonably certain, that judgment necessarily stays with the client.

The cleaner arrangement: let the client run it

On an attest engagement, the tidiest answer is frequently that the firm does not perform the abstraction at all. The client uploads their own documents, reviews the extracted fields against the citations, and owns the resulting schedule. The firm then audits that schedule the way it audits anything else.

This removes the nonattest analysis rather than managing it. There is no service to document, no SKE individual to designate for work the firm performed, and no argument at peer review about whether oversight was meaningful. It also tends to be faster, because the client is not waiting on firm capacity in February. That is the workflow behind our lease abstraction for CPA firms page: extraction on one side of the line, judgment and audit on the other.

What if there is no attest relationship?

Then most of this falls away. For a consulting only client, or a client where the firm performs no attest engagement, lease abstraction is ordinary advisory work and the independence rules in ET 1.295 do not apply. Firms doing lease implementation as a standalone service line often do the abstraction, build the schedules, and hand over a completed model.

Compilation and review engagements sit in between and still require independence for reviews, so the same general requirements apply. Compilations permit a lack of independence provided it is disclosed, which is a business decision rather than a technical bar. Check the specific engagement before assuming the rules are looser than they are.

Finding the leases is a separate problem

Independence questions assume you know what the population is. Often nobody does. Base leases are usually easy to collect; what goes missing are amendments that changed a term years after signing, month to month arrangements nobody documented, and embedded leases buried inside logistics, equipment, and managed service agreements that were never labeled as leases.

A workable sweep pairs the accounts payable ledger against rent expense and fixed asset accounts, then requests every contract above a dollar threshold rather than every contract someone filed under "leases". For clients whose agreements are scattered across shared drives, email, and three generations of document systems, the practical prerequisite is being able to search every internal system at once instead of asking department heads to remember. Completeness testing fails on documents nobody knew existed far more often than it fails on documents someone read incorrectly. Our guide to embedded leases under ASC 842 covers where they typically hide.

A short checklist before the work starts

• Confirm whether an attest relationship exists, and which type.
• Name a specific SKE individual and satisfy yourself they could actually evaluate a lease schedule.
• Document the service precisely in writing, before it begins.
• Draw the reading versus concluding line explicitly, in the engagement letter.
• Keep the option, discount rate, and lease identification conclusions with the client, in their own documentation.
• Prefer a deliverable that is traceable to source pages, so oversight is evidence rather than a signature.

Handled this way, lease abstraction is a service most firms can offer without much drama. The risk is not in the reading. It is in letting the reading quietly turn into deciding, and discovering the difference at peer review. For the volume side of the problem, see how long lease abstraction takes, and for the tools themselves, best lease abstraction software.