Lease accounting is deceptively complex. Here's what AI handles well — and the judgment calls that still need a human.
ASC 842 has been fully effective for a while now, but lease accounting is still generating hard questions.
Part of that is the ongoing nature of the work — new leases, modifications, renewals, terminations, changes in lease terms. Unlike a one-time adoption, lease accounting is a continuous compliance obligation. New transactions keep coming, and each one requires its own analysis.
Part of it is that 842 is genuinely more complex than it looks. The lessee model seems straightforward on the surface — identify the lease, classify it, record the right-of-use asset and liability — but the details are where things get complicated. Variable payments, purchase options, related party leases, embedded leases in service contracts, short-term lease elections. Any of these can change the analysis, and not all of them are obvious from a quick read of the contract.
I spent time at Netgain, a company that built lease accounting software and helped companies through 842 adoptions. I've seen the full range of what goes wrong — and where better tools, including AI, make a real difference.
What ASC 842 compliance actually involves
To understand where AI helps, it helps to be clear about what 842 compliance actually requires on an ongoing basis. It's more than just running the amortization schedule.
Lease identification. Before you can account for a lease, you have to identify whether an arrangement contains a lease at all. Under 842, this means evaluating whether a contract conveys the right to control the use of an identified asset for a period of time in exchange for consideration. This analysis is more nuanced than it sounds — especially for contracts that are primarily service arrangements but have lease components embedded in them.
Lease classification. Lessees classify leases as finance or operating, depending on criteria that relate to the lease term relative to the asset's useful life, the present value of lease payments relative to fair value, specialized nature of the asset, and other factors. The classification affects how costs are presented in the income statement.
Initial measurement. The right-of-use asset and lease liability are measured at commencement. This requires identifying all lease payments — fixed payments, variable payments based on an index or rate, amounts likely to be owed under residual value guarantees, and payments for purchase or renewal options if exercise is reasonably certain. Getting this wrong at commencement creates errors that carry forward for the life of the lease.
Subsequent measurement. The liability is remeasured when certain triggering events occur — a lease modification, a change in the assessment of exercise of an option, a change in the amount expected to be owed under a residual value guarantee, a change in payments based on an index or rate. Each remeasurement requires its own analysis and documentation.
Modifications. Lease modifications are accounted for differently depending on whether the modification grants an additional right of use not included in the original lease and whether the consideration is commensurate with the standalone price. A modification that qualifies as a separate lease is accounted for prospectively. One that doesn't is accounted for through a remeasurement of the existing lease. Getting this right requires reading the modification carefully and understanding what changed.
Where AI genuinely helps with 842
Lease identification analysis
The "does this contract contain a lease" question is one of the most consistently underestimated parts of 842 compliance. It comes up constantly — in new vendor contracts, in IT arrangements, in real estate service agreements — and the analysis requires working through the specific language of the contract against the 842 criteria.
AI is well-suited to this. Given the contract terms, it can walk through the identification criteria systematically, flag the parts of the analysis that require judgment, and help you document the conclusion. For straightforward arrangements, this can be handled quickly. For more complex ones, the AI gives you a structured starting point.
Explaining the standard in plain language
One of the most practical uses of AI in lease accounting is explaining the guidance to people who aren't CPAs — operations teams, legal teams, finance business partners who need to understand why a particular arrangement is being treated a certain way.
842 is dense. AI can translate the guidance into plain language, explain the accounting in terms of the business facts, and help you communicate the conclusion without requiring everyone to read the codification themselves.
Modification analysis
Lease modifications are a recurring area of complexity. The question of whether a modification is a separate lease or a modification of the existing lease requires careful analysis of both the facts and the guidance.
AI that's been built with 842 knowledge can work through that analysis with you — walking through the criteria, applying them to the specific modification terms, and helping you document the conclusion and the accounting treatment that follows.
Research on specific questions
842 generates a lot of specific questions: How do we handle a lease where we're the lessor? What's the right treatment for a lease with a purchase option that's not reasonably certain? How do we account for a lease of a floor in a building we partially own?
AI can help you work through these systematically — finding the relevant guidance, surfacing the applicable examples, and helping you structure the analysis. It's faster than navigating the codification cold, especially for questions you haven't encountered before.
Where AI still falls short
The spreadsheet problem
A lot of companies are still managing leases in spreadsheets. If you're in that situation, AI can help you think through the accounting, but it can't audit your spreadsheet for you.
I've seen firsthand what happens when spreadsheet-based lease accounting goes wrong — missing components, calculation errors, assumptions that were reasonable at adoption but haven't been updated. The accounting can look right for years until a migration or an audit surfaces a problem. AI doesn't fix that structural issue.
The honest recommendation here is to use dedicated lease accounting software for the calculations. Tools like NetLease are purpose-built to handle the amortization schedules, remeasurement triggers, and disclosure rollforwards correctly every time. That's not something you want to rely on a spreadsheet — or an AI — to get right. The calculation work needs a system with proper controls.
Where AI fits alongside that software is in the analysis and documentation work that happens around the calculations: identifying whether a new contract contains a lease, determining the right classification, working through a modification, answering a one-off research question. That's the judgment layer — and it's where AI actually helps.
Reasonably certain assessments
Determining whether a renewal option, purchase option, or early termination option is "reasonably certain" to be exercised is one of the most judgment-intensive parts of 842. It requires understanding the economics, the business intent, the history, and the facts on the ground — none of which AI can know from the standard alone.
AI can tell you what factors to consider. It can't make the assessment for you, and it can't know whether your company is actually likely to exercise a five-year renewal option in a lease that's three years from expiration.
Related party leases
Related party leases require disclosure of the terms and purpose of the arrangement, and the accounting often requires careful analysis of whether the arrangement is structured to reflect market terms. This is an area where the facts outside the lease document matter a lot — and where AI is limited to what you tell it.
A practical framework for AI-assisted 842 work
The right way to think about the 842 toolstack is: dedicated lease accounting software for the calculations, AI for the analysis and documentation around them. They're solving different problems.
Your lease accounting software — something like NetLease — handles the amortization, the remeasurement entries, the rollforward schedules, and the disclosure support. It's built for that work, and it gets the math right every time. You still need to set it up correctly and feed it the right inputs, but the calculation engine isn't where you should be improvising.
AI fits into the workflow around that software:
For new leases: work through the identification and classification analysis before the lease goes into the system. AI can help you document why the arrangement is or isn't a lease, how you classified it, and what the key judgment calls were.
For modifications: use AI to work through the accounting treatment question — is this a separate lease or a remeasurement? — before you make any entries. The modification analysis needs to be documented, and getting it right up front avoids remeasurement errors downstream.
For research questions: use AI to find the relevant guidance, work through the applicable examples, and structure the analysis before you write the memo.
For everything: verify the citations, review the application to your specific facts, and make sure the judgment calls are documented — not just the conclusions.
Lease accounting is ongoing work. The goal is a process that's consistent, documented, and defensible across every period. The software handles the calculations. AI helps you get the analysis and documentation right. Together they cover more ground than either one does alone.
Lease accounting is one of Gaapio's core areas of focus — identification analysis, modification treatment, ongoing remeasurement. If 842 questions come up regularly in your work, it's worth seeing how it handles a specific scenario.