For any finance professional dealing with fixed assets, goodwill, or intangible assets, the annual impairment test under IAS 36 is a recurring source of friction. The process is deceptively complex: you must identify cash-generating units (CGUs), allocate goodwill, build discounted cash flow (DCF) models, determine appropriate discount rates, perform sensitivity analyses, and document every assumption for the auditors. One misplaced cell reference or an inconsistent growth rate can trigger a restatement or, worse, a qualified audit opinion. The time cost is immense—senior analysts and controllers routinely spend 40 to 80 hours per reporting cycle just on impairment testing for a mid-sized group.
This pain is compounded by the sheer volume of data that must be reconciled. Budget figures, five-year forecasts, historical growth rates, industry benchmarks, and recent transaction multiples all need to be pulled from disparate systems and synthesized into a coherent model. The result is a spreadsheet-heavy process that is error-prone, difficult to review, and nearly impossible to replicate consistently across subsidiaries. Most teams end up with dozens of versions of the same model, each with slightly different assumptions, making the audit trail a nightmare to reconstruct.
AI tools—particularly large language models like Claude—can eliminate the drudgery while improving accuracy. By acting as your structured reasoning partner, an AI can ingest your financial data, apply IAS 36 logic, generate draft calculations, and flag inconsistencies before they reach the audit committee. The key is learning how to prompt the AI with enough context and structure to produce reliable, auditable outputs. This post provides two ready-to-use prompt templates that will transform your impairment testing workflow.
Why Traditional Impairment Testing Fails Under Pressure
The fundamental challenge with IAS 36 is that it requires judgment calls at every step. What is the appropriate pre-tax discount rate? Should you use a perpetuity growth rate of 2% or 2.5%? Which CGU allocation best reflects how management monitors the business? These questions cannot be answered by a formula alone—they require a defensible narrative. Yet the manual process of building that narrative is where most teams get stuck. They spend days formatting spreadsheets instead of thinking critically about assumptions. An AI prompt, when properly structured, forces you to articulate your logic upfront and then executes the mechanical calculations in seconds, freeing you to focus on the judgment elements.
First, read these files completely before responding:
[segment_financials_2025.xlsx] — Contains the P&L, balance sheet, and cash flow data for EMEA Industrial for FY2025, including revenue, EBITDA, capex, and working capital movements.
[audit_prior_year_impairment.pdf] — The signed impairment working paper from FY2024, including the discount rate calculation, growth assumptions, and sensitivity tables.
[cgU_allocation_schedule.xlsx] — The current goodwill allocation across CGUs, with the EMEA Industrial segment broken into three CGUs: Germany, France, and Benelux.
Here is a reference for what I want to achieve:
I have attached a sample impairment model from a peer company in the same industry (industrial manufacturing). It shows how they structure the value-in-use calculation, including the WACC build-up, terminal value formula, and sensitivity matrix.
Here’s what makes this reference work:
– The DCF model uses a 5-year explicit forecast period, followed by a perpetuity growth rate of 2.0% to 2.5%.
– The discount rate is calculated using the CAPM with a local risk-free rate, equity risk premium, and a beta derived from comparable companies.
– Sensitivity is shown for revenue growth (-5% to +5%) and discount rate (+/- 100 bps).
– Every assumption has a footnote explaining its source (e.g., “Bloomberg as of 31 Dec 2025”).
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: A complete value-in-use calculation for each of the three CGUs, presented in a table with explicit formulas. Total output should be approximately 2-3 pages.
Recipient’s reaction: The CFO should be able to review the draft and immediately identify which assumptions need further discussion. The auditors should see a clear audit trail for every input.
Does NOT sound like: A generic textbook example. It must use our actual segment data and reflect the specific risk profile of industrial manufacturing in Europe.
Success means: I can copy the output directly into our impairment working paper template, requiring only minor formatting adjustments.
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.
How This Prompt Changes the Game
The prompt above works because it forces the AI to operate within a strict logical framework. Notice the explicit instructions: read the files first, extract patterns from a reference model, and confirm the execution plan before producing anything. This prevents the AI from jumping to generic answers. For a controller running the impairment test for a multi-segment group, this structure ensures that every assumption is grounded in actual data, not AI hallucination. The “success brief” section is particularly powerful because it defines exactly what the output must achieve—auditability and CFO-readiness—rather than just listing tasks.
One practical tip: always include a “prior year working paper” as a reference file. The AI will use it to infer your organization’s preferred discount rate methodology, typical sensitivity ranges, and documentation style. This dramatically reduces the amount of manual editing you need to do after the output is generated. In our tests, teams that provided a prior-year audit file saw a 60% reduction in rework time compared to those who only uploaded current-year data.
First, read these files completely before responding:
[APAC_CGU_model_output.xlsx] — The completed value-in-use calculation for the APAC CGU, including base-case headroom of EUR 12.4 million.
[market_risk_premium_report.pdf] — Damodaran’s January 2026 country risk premium update for China, India, and Southeast Asia.
[board_presentation_template.pptx] — The standard format used for presenting impairment conclusions to the audit committee.
Here is a reference for what I want to achieve:
The attached file “sensitivity_template_v3.xlsx” shows how our external auditors (PwC) require the sensitivity analysis to be structured: a three-way table with discount rate on rows, terminal growth rate on columns, and headroom / impairment in the cells. It also includes a “reasonably possible change” scenario that is required by IAS 36.36.
Here’s what makes this reference work:
– The sensitivity table uses increments of 0.5% for both discount rate and growth rate.
– The “reasonably possible change” scenario is presented as a separate table showing the impact of a 1% decline in revenue growth combined with a 50 bps increase in discount rate.
– Each scenario includes a clear conclusion: “No impairment” or “Impairment of EUR X” with the triggering assumption highlighted in bold.
– The memo includes a paragraph explaining management’s view on the likelihood of each scenario.
Here’s what I need for my version / SUCCESS BRIEF:
Type of output + length: A one-page memo (300-400 words) plus a sensitivity table. The memo must be ready to copy into the board presentation slide.
Recipient’s reaction: The audit committee should be able to read the conclusion in 60 seconds and feel confident that no material impairment risk exists. If a scenario shows impairment, the trigger must be immediately obvious.
Does NOT sound like: A compliance checkbox exercise. It must read as a thoughtful, risk-aware analysis that acknowledges uncertainty but provides a clear management view.
Success means: The external auditors accept the memo as sufficient audit evidence for the APAC CGU, and no follow-up questions are raised.
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.
Practical Next Steps for Your Team
The two prompts above cover the most critical phases of any IAS 36 impairment test: the initial value-in-use calculation and the sensitivity analysis with conclusion memo. To get the most out of this approach, start by running the first prompt at the beginning of your impairment testing cycle. Use it to generate a draft calculation for your largest or most complex CGU. Review the output critically—the AI will make mistakes if your data is inconsistent or if the reference files contain errors. Treat the AI output as a highly competent first draft, not a final product. Then, run the second prompt after you have validated the base-case numbers, using the refined model as input.
One common pitfall is uploading incomplete or poorly organized data. Before you start, ensure your financial files are clean: no merged cells, no hidden rows, no inconsistent currency formats. The AI reads the data as it appears, so a spreadsheet with “EUR” in one column and “€” in another will produce garbled output. A few minutes of data cleaning upfront will save hours of editing later. Also, consider running both prompts for your smallest CGU first as a proof of concept. Once you see the quality of output, you will have the confidence to scale the approach across all reporting segments.
Finally, remember that the AI cannot replace your professional judgment. The discount rate, growth assumptions, and CGU boundaries remain your responsibility. What the AI does is remove the mechanical friction, enforce consistency, and provide a clear audit trail. Use the time you save to focus on the strategic questions: Are our CGUs correctly defined? Is our goodwill allocation still appropriate after the latest acquisition? That is where the real value lies.
Published on 7 June 2026 on growwithgpt.com