The transition to IFRS 16 was supposed to bring transparency to off-balance-sheet leasing activity. Instead, for many finance teams, it has brought a permanent administrative burden. Every quarter, your analysts are likely wrestling with spreadsheets containing hundreds of lease lines, manually recalculating right-of-use (ROU) assets and lease liabilities, adjusting for modifications, reassessments, and terminations. The risk of material misstatement is high: a single incorrect discount rate or a missed renewal option can distort your balance sheet and income statement. This is not just a technical headache—it is a direct threat to audit confidence, investor trust, and the efficiency of your close cycle.
This is where a purpose-built AI workflow changes the game. Instead of using generic chatbots that give you theoretical advice, you can now deploy a structured prompt system that acts like a senior technical accountant. This AI tool does not merely answer questions—it executes. It reads your lease portfolio, applies the IFRS 16 standard to each contract, calculates the amortization schedules, and produces journal entries in a format your ERP system understands. The result is a reduction in manual effort by over 80%, a significant drop in calculation errors, and a fully auditable trail from source contract to final disclosure.
However, the key to making this work is not the AI model itself—it is the quality of the instructions you give it. A vague prompt like “calculate my leases” will yield a generic, unusable response. What you need is a precise, structured prompt that defines the exact task, the input files, the output format, and the success criteria. Below, I have provided a two-part prompt template that you can adapt to your specific portfolio. The first prompt handles the initial bulk calculation of your lease population. The second prompt handles the trickier, ongoing task of lease modifications and reassessments.
Why Most Automation Attempts Fail (And How to Fix It)
Before we dive into the prompts, it is worth understanding why most finance teams struggle to automate this process. The primary reason is that they treat AI like a calculator rather than a junior accountant. They ask for a single number without providing the context of the lease agreement, the payment schedule, or the incremental borrowing rate (IBR). This leads to hallucinations—the AI makes up plausible-looking numbers that are completely wrong. The second reason is a lack of a defined output schema. If you do not tell the AI exactly which columns to produce (e.g., Period, Payment, Interest Expense, Depreciation, Closing Liability), it will invent its own format, making it impossible to import into your general ledger.
The solution is a two-stage approach. First, you create a “context file” that contains your company’s accounting policies, your standard discount rate assumptions, and a sample of correctly calculated leases. This file acts as the ground truth for the AI. Second, you use the structured prompt below to force the AI to read that context file, ask clarifying questions if needed, and then execute the calculation in a strict, step-by-step manner. This method turns the AI from a passive text generator into a reliable, rule-based processing engine.
Let’s look at the first prompt—the one you will use for your initial portfolio transition or your quarterly re-calculation of all active leases.
First, read these files completely before responding:
lease_portfolio_raw.csv — Contains all lease contracts with fields: Lease_ID, Commencement_Date, Lease_Term_Months, Payment_Frequency, Annual_Payment_Amount, Renewal_Option_Years, Purchase_Option_Flag, Residual_Value_Guarantee.
company_ibR_policy.md — Contains our incremental borrowing rate curve segmented by lease term (1-3 years, 4-6 years, 7-10 years) and by currency.
ifrs16_calculation_sample.xlsx — Contains a manually verified example of 5 leases with the correct amortization schedule and journal entries.
Here is a reference for what I want to achieve:
The reference file (ifrs16_calculation_sample.xlsx) demonstrates the exact layout of the output: a separate tab for the amortization schedule, a tab for the journal entry summary, and a tab for the disclosure note text.
Here’s what makes this reference work:
The amortization schedule uses a monthly period granularity, even for annual payments, to allow for straight-line depreciation of the ROU asset.
The discount rate is applied to the remaining liability balance, not the original principal.
The journal entries are split into three lines: Dr Depreciation, Dr Interest Expense, Cr Cash, and Cr Lease Liability.
The disclosure note uses the exact wording required by IFRS 16 paragraph 53, including the maturity analysis table.
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: A structured Excel workbook with three sheets (Amortization, Journal_Entries, Disclosures) covering all 250 leases in the portfolio.
Recipient’s reaction: The external auditor should be able to trace any number on the balance sheet back to a specific lease line in the workbook without asking a single follow-up question.
Does NOT sound like: A generic summary or a list of recommendations. It must be a completed, populated workbook, not a template.
Success means: The total lease liability at the reporting date matches my manual calculation within a 0.01% tolerance, and the output is formatted for direct upload into SAP.
My context file contains my standards, constraints, audience. Read it fully before starting.
DO NOT start executing yet. Ask clarifying questions first.
Specifically, ask me: (1) whether to use the short-term lease exemption for leases under 12 months, (2) whether to include non-lease components (e.g., maintenance) in the payment, and (3) what exchange rate to use for the reporting date.
Give me your execution plan (5 steps max) before you begin.
The prompt above is designed to be rigorous. Notice how it forces the AI to read specific files before doing anything. It also explicitly asks for a “success criteria” that is measurable—a 0.01% tolerance against a manual check. This is critical because it prevents the AI from “rubber-stamping” its own incorrect output. The final instruction to “ask clarifying questions” is not a formality. In practice, this single line often prevents hours of rework, because it surfaces ambiguities in your lease data (e.g., is a payment inclusive or exclusive of VAT?) before the calculation runs.
Once you have the initial calculation done, the real work begins. Lease portfolios are not static. Every month, you have modifications: an extension of a lease term, a renegotiation of rent, or an early termination. Each of these events requires a recalculation of the liability and an adjustment to the ROU asset. This is where the second prompt comes in. It is designed to handle the “delta” rather than the entire portfolio, which saves significant processing time and reduces the risk of introducing errors into already-verified leases.
First, read these files completely before responding:
lease_modifications_2026_q3.csv — Contains the following fields: Lease_ID, Modification_Type (Extension, Termination, Rent_Change), New_End_Date, New_Annual_Payment, Effective_Date.
current_lease_balances.xlsx — Contains the opening ROU asset balance and lease liability balance for each lease as of the last reporting period.
ifrs16_modification_rules.md — Contains the technical guidance on how to treat each modification type (e.g., extension increases both ROU and liability; termination reduces both proportionally).
Here is a reference for what I want to achieve:
A reference file (modification_workbook_example.xlsx) shows how a single modification was processed: the old liability was derecognized, the new liability was measured at the revised discount rate, and the difference was posted to the ROU asset.
Here’s what makes this reference work:
The discount rate is updated to the current market rate at the modification date, not the original rate.
The ROU asset adjustment is calculated as the difference between the old and new liability, not as a plug figure.
The journal entry includes a specific line item for the gain/loss on modification when the ROU asset is reduced to zero.
The effective date of the modification is used as the start date for the new amortization schedule.
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: A concise exception report (max 3 pages) listing each modified lease, the adjustment amount, and the proposed journal entry. Plus a separate CSV file with the updated amortization schedule for just the modified leases.
Recipient’s reaction: The CFO should be able to review the total P&L impact of the modifications in under 10 minutes without opening the lease spreadsheet.
Does NOT sound like: A narrative explanation of what IFRS 16 is. It must be a specific, actionable set of adjustments.
Success means: The sum of all modification adjustments equals the change in the total lease liability roll-forward schedule, and every modified lease has a clear audit trail from the modification event to the new liability balance.
My context file contains my standards, constraints, audience. Read it fully before starting.
DO NOT start executing yet. Ask clarifying questions first.
Specifically ask me: (1) whether any of the modifications involve a change in the lease term that triggers a reassessment of the purchase option, (2) whether the company has elected to use the portfolio approach for discount rates, and (3) whether the modifications should be weighted average or specific identification for the discount rate.
Give me your execution plan (5 steps max) before you begin.
When you use these two prompts in tandem, you create a powerful, closed-loop process. The first prompt builds your baseline. The second prompt maintains it. The practical tip here is to not treat these prompts as static scripts. They are living documents. After the first successful run, review the output and refine the “Here’s what makes this reference work” section. For example, you may discover that your auditor prefers a different rounding convention or that your ERP system requires a specific account code format. Add those details to the prompt context file so the AI learns your preferences over time.
What to try next? I recommend starting with a pilot group of 10 leases that you have already calculated manually. Run the first prompt against that subset and compare the AI output to your manual numbers. This will validate your context files and prompt structure before you scale to the full portfolio. Once validated, you can schedule this prompt to run at the beginning of every month, feeding the output directly into your consolidation software. The goal is to reduce your lease accounting close activities from three days to three hours.
This approach does not eliminate the need for a skilled accountant—it elevates them. Instead of spending hours on data entry and formula checks, your team can focus on reviewing exceptions, negotiating better lease terms, and analyzing the impact of leasing on your financial ratios. The AI handles the arithmetic; you handle the judgment. That is the true promise of automation in finance.
Published on 16 August 2026 on growwithgpt.com