AI Scenario Planning
Executive Summary
Key Takeaways
- ✓ AI-assisted scenario planning uses generative AI to draft a wider range of plausible scenario variations around a base case than a modeller might generate manually, accelerating scenario ideation.
- ✓ An AI-drafted scenario must still be checked for internal consistency, whether its assumptions are mutually compatible, before being used, since a language model can draft a scenario whose individual assumptions do not actually cohere together.
- ✓ Probability weighting across a set of scenarios remains a judgement exercise; generative AI can describe a scenario's plausibility narratively, but assigning it a specific probability weight requires the same commercial judgement any assumption requires.
- ✓ The mechanical breadth AI provides, generating more scenario variations faster, is valuable only if paired with the same consistency checking and probability judgement any scenario planning process requires, whether AI-assisted or not.
- ✓ This guide's scope is scenario variation generation specifically; the driver ranking and range-setting discipline underlying sensitivity analysis is addressed separately in AI Sensitivity Analysis.
Objective¶
This guide sets out generative AI's role in scenario planning, within AI for FP&A and the broader AI Financial Modelling & Artificial Intelligence in Finance domain.
Where Generative AI Adds Value¶
Generative AI can draft a wider range of plausible scenario variations around a base case than a modeller might generate manually within the same time, upside and downside variations across several dimensions at once, accelerating the ideation stage of scenario construction described more generally in Generative AI in Financial Modelling.
The Consistency Check¶
An AI-drafted scenario must be checked for internal consistency before use: whether its individual assumptions actually cohere together as a plausible combined state of the world. A language model can draft a scenario where each individual assumption reads plausibly in isolation, a lower interest rate alongside a specific volume assumption, for instance, while the combination does not actually reflect a coherent, internally consistent economic scenario. This check is a modeller's responsibility and is not automatically performed by the drafting process itself.
Probability Weighting Remains Judgement¶
Assigning a probability weight to a scenario, how likely is this specific combination of assumptions relative to the base case and other scenarios, remains a judgement exercise. Generative AI can describe a scenario's qualitative plausibility in narrative terms, but a specific probability weight used to combine scenarios into an expected value or a weighted range requires the same commercial judgement any such weighting has always required, informed by AI-drafted scenario breadth rather than replaced by it.
Why Breadth Alone Is Not the Goal¶
Generating more scenario variations faster is valuable only if each variation is paired with the same consistency checking and probability judgement any scenario planning process requires. A larger set of unchecked, unweighted scenarios does not itself improve decision quality; it is the reviewed, weighted subset of genuinely plausible and internally consistent scenarios that supports a decision.
Common Construction Pitfalls¶
Using an AI-drafted scenario without a consistency check. An unchecked scenario can combine individually plausible but mutually incompatible assumptions, understating or overstating the scenario's actual likelihood.
Treating scenario breadth as a substitute for probability judgement. A wide set of AI-drafted scenarios still requires the same weighting judgement a smaller, manually generated set would require.
Skipping documentation of why a scenario was accepted or rejected. Recording the consistency check and weighting rationale for each AI-drafted scenario preserves the audit trail a material decision typically requires.
Recommended Practices¶
- Use generative AI to broaden the initial set of scenario variations considered, not to finalise them.
- Check every AI-drafted scenario for internal consistency before use.
- Apply the same commercial judgement to probability weighting regardless of how the scenario was drafted.
- Document the review and weighting rationale for each scenario retained for use.
Continue Reading¶
Related Pillars¶
Related Technical Guides¶
Related Comparisons¶
How OXXON tests thisRun a free structural check with FMAE
Frequently Asked Questions
How does generative AI assist scenario planning?
By drafting a wider range of plausible scenario variations around a base case than a modeller might generate manually, accelerating the ideation stage of scenario construction, which the modeller then reviews and refines.
Should an AI-drafted scenario be used without further review?
No. It must still be checked for internal consistency, whether its individual assumptions actually cohere together, since a language model can draft a scenario whose assumptions read plausibly in isolation but do not fit together as a coherent whole.
Can generative AI assign probability weights to scenarios?
Not reliably as a substitute for judgement. Probability weighting across a set of scenarios remains a judgement exercise; generative AI can describe a scenario's plausibility narratively, but assigning it a specific probability weight requires the same commercial judgement any assumption requires.
Does AI scenario planning replace sensitivity analysis?
No, the two are related but distinct. This guide addresses scenario variation generation specifically; the driver ranking and range-setting discipline underlying sensitivity analysis is addressed separately in AI Sensitivity Analysis.
Related Articles
AI Financial Modelling & Artificial Intelligence in Finance
AI financial modelling is the application of machine learning and generative AI techniques within the financial modelling process itself, driver identification, construction assistance, scenario generation, and narrative drafting, while artificial intelligence in finance is the broader application of those same technique categories across the finance function generally. This page is the hub for the Knowledge Centre's AI financial modelling content: the foundational distinction between machine learning, natural language processing, and generative AI; how AI accelerates modelling construction without replacing the auditable calculation layer beneath it; a staged framework for adopting AI reliably; enterprise applications across FP&A, forecasting, valuation, and investment analysis; governance and risk practice; and the institutional best practice synthesis this domain builds toward.
AI for FP&A
FP&A is one of the finance functions where AI adoption has progressed furthest, spanning budgeting, forecasting, variance analysis, and management reporting. This guide maps AI application across the FP&A cycle, distinguishes tasks where machine learning-driven prediction is well matched from tasks better suited to generative AI drafting, and sets the frame for the more specific budgeting, forecasting, and analysis guides that follow it in this domain.
AI Sensitivity Analysis
AI-assisted sensitivity analysis uses machine learning to help rank which drivers most influence a model's output across historical data, directing attention to the variables worth testing most rigorously, and generative AI to draft commentary explaining sensitivity results. This guide sets out that role and why the specific sensitivity ranges tested against each driver should remain a modeller's defined, documented, and auditable input rather than an AI-generated range.
AI Forecasting Models
AI forecasting models use machine learning to predict a future value from patterns learned in historical data, complementing traditional driver-based forecasting rather than replacing it. This guide sets out the training data requirements a machine learning forecast depends on, how its accuracy should be measured and monitored over time, and the specific forecasting tasks where a machine learning approach adds genuine value over a traditional driver-based model.
Generative AI in Financial Modelling
Generative AI, large language models applied to drafting and language tasks, has a specific and bounded role in financial modelling: accelerating structure, formatting, and narrative drafting, not producing verified numerical output. This guide sets out that role in detail, the specific failure modes generative AI introduces into a modelling workflow, hallucinated figures, plausible-but-incorrect formula logic, and unverifiable citations, and the concrete review practices that contain each failure mode.
AI Scenario Planning vs. Traditional Scenario Planning
AI-assisted and traditional scenario planning both aim to construct a set of plausible future states around a base case, but differ in how scenario variations are generated and how quickly a broader set can be produced. This comparison sets out those differences and confirms that the consistency checking and probability weighting responsibilities remain the same regardless of which approach generated the initial scenario set.