IFRS 16 turned lease accounting into a data engineering problem. Every operating lease that used to sit quietly in the notes now has to be capitalized: a right-of-use asset, a lease liability, a discount rate, a remeasurement schedule, and a stream of interest and depreciation entries that must reconcile to the cent across the full lease term. For a company with a handful of property leases, a spreadsheet works. For a company with hundreds of leases across vehicles, equipment, cell towers, and retail locations, each with renewal options, indexation clauses, and modification events, the spreadsheet becomes a liability of its own.
The friction is not the math. The math is well defined. The friction is everything around it: extracting commencement dates and payment schedules from PDFs, deciding whether a renewal option is “reasonably certain” to be exercised, handling a CPI escalation mid-term, catching the lease that was modified in June but only surfaced in the October reconciliation, and producing an audit-ready disclosure note that ties to the general ledger. Controllers spend weeks on this every reporting cycle, and the errors that slip through tend to be expensive ones.
This is where a structured AI workflow changes the economics. Claude, given the right context files and a disciplined prompt, can parse lease documents at scale, build and stress-test amortization schedules, flag judgmental assumptions for human review, and draft the disclosure notes and audit support memos that normally consume the last two weeks of the close. It does not replace the accountant’s judgment on “reasonably certain” — it forces that judgment to be documented, consistent, and reviewable.
Where AI fits in the IFRS 16 workflow
The practical gains come in three places: extraction (turning lease PDFs into structured data), calculation (building schedules that survive auditor scrutiny), and communication (memos and disclosures that explain the numbers). The two prompts below cover the highest-leverage versions of these tasks.
Prompt 1: Lease data extraction and schedule build
Use this when you are onboarding a lease portfolio into your calculation engine, or when you inherit a messy lease register and need to rebuild it from source documents.
First, read these files completely before responding:
[lease_register.xlsx] — current lease register with lease IDs, counterparties, asset types, and any known gaps
[ifrs16_accounting_policy.md] — our accounting policy, including discount rate methodology, low-value and short-term exemptions, and remeasurement triggers
[chart_of_accounts.csv] — account codes for ROU asset, lease liability (current and non-current), interest expense, and depreciation
Here is a reference for what I want to achieve:
[Upload a completed lease schedule from a prior period as markdown, showing one lease fully worked: commencement date, term, payments, discount rate, liability rollforward, ROU asset rollforward, and journal entries]
Here’s what makes this reference work:
It shows every input field labeled exactly as our system expects, it reconciles the liability rollforward to the ROU rollforward, it separates contractual payments from non-lease components, and it documents the discount rate source in a single line rather than burying it in a footnote.
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: Structured table per lease plus a portfolio summary, one row per lease, all fields populated or explicitly marked as requiring judgment
Recipient’s reaction: The controller should be able to spot-check any lease in under two minutes and trust the numbers without re-deriving them
Does NOT sound like: A generic accounting summary, a narrative essay, or a spreadsheet with unexplained blanks
Success means: Every lease in the register is matched to a source document, every schedule foots, and the only open items are documented judgment calls on renewal options and discount rates
My context file contains my standards, constraints, audience. Read it fully before starting.
DO NOT start executing yet. Ask clarifying questions first.
Give me your execution plan (5 steps max) before you begin.
Note the deliberate constraint: the prompt forces the model to flag judgment calls rather than silently making them. That single instruction is what keeps the output defensible in front of an auditor.
Prompt 2: Remeasurement, disclosure, and audit memo
Use this at period end, or whenever a lease modification, indexation reset, or term reassessment occurs. This is the task that historically eats the most controller time.
First, read these files completely before responding:
[lease_schedule_prior_period.xlsx] — prior period schedules with liability and ROU balances per lease
[modification_events_log.md] — list of modifications, indexation resets, and term reassessments during the period, with dates and source documents
[disclosure_template_prior_year.md] — last year’s published IFRS 16 note, including maturity analysis table format
Here is a reference for what I want to achieve:
[Upload a prior-period audit memo as markdown covering one remeasurement, showing the trigger, the revised cash flows, the revised discount rate, the adjustment to liability and ROU, and the journal entry]
Here’s what makes this reference work:
It states the trigger and the standard paragraph it relies on, quantifies the adjustment before and after, shows the journal entry in debit and credit form, and separates facts from assumptions in clearly labeled sections.
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: One remeasurement memo per event plus a consolidated disclosure note with maturity analysis, total under 1,500 words unless the event count requires more
Recipient’s reaction: The audit partner should read it once and sign off, and the CFO should be able to explain the P&L impact in one sentence
Does NOT sound like: A textbook explanation of IFRS 16, a hedge, or a memo that leaves the discount rate unexplained
Success means: Every remeasurement ties to a source document, the disclosure note agrees to the lease liability rollforward, and no open questions remain in the audit query log
My context file contains my standards, constraints, audience. Read it fully before starting.
DO NOT start executing yet. Ask clarifying questions first.
Give me your execution plan (5 steps max) before you begin.
One practical tip before you deploy either prompt: run them on a closed prior period first, where you already know the correct answer. Compare the AI output to your audited numbers line by line. Anywhere it diverges, the gap is almost always a missing context file rather than a model failure — add the policy extract or the prior schedule and rerun. Once the prompts reproduce a known-good period, they are ready for live use.
Start with Prompt 1 on your ten most complex leases. If the extraction and schedules hold up, expand to the full portfolio and layer in Prompt 2 at the next close. The goal is not to remove the accountant from the loop — it is to move the accountant from building schedules to reviewing judgment, which is where their expertise actually creates value.
Published on 16 September 2026 on growwithgpt.com