Foundation Model
Executive Summary
Key Takeaways
- ✓ A foundation model is a large-scale AI model trained on broad, general-purpose data and designed to be adapted to specific tasks through fine-tuning or prompting, rather than trained from scratch for each new application.
- ✓ Most generative AI tools used in finance today are built on a general-purpose foundation model rather than a model trained specifically and exclusively on financial data.
- ✓ Because a foundation model's underlying training data is general-purpose, its knowledge of finance-specific conventions, terminology, and edge cases should not be assumed to be complete or current without verification.
- ✓ Adapting a foundation model to a finance-specific task typically relies on prompt engineering or supplying task-specific source data at the point of use, rather than on the model having been trained specifically for finance.
Definition¶
A foundation model is a large-scale AI model, typically a large language model, trained on broad, general-purpose data and designed to be subsequently adapted to specific tasks through fine-tuning or prompting, rather than trained from scratch for each new application.
Why This Matters for Finance-Specific Use¶
Most generative AI tools used in finance today are built on top of a general-purpose foundation model rather than a model trained specifically and exclusively on financial data. Because its underlying training data is general-purpose, a foundation model's knowledge of finance-specific conventions, terminology, and edge cases should not be assumed complete or current without independent verification, particularly for anything jurisdiction- or entity-specific.
Adapting a Foundation Model to Finance¶
Adapting a foundation model to a finance-specific task typically relies on prompt engineering, or on supplying task-specific source data at the point of use, rather than on the underlying model having been retrained specifically for finance.
Continue Reading¶
Related Pillars¶
Related Glossary¶
How OXXON tests thisRun a free structural check with FMAE
Frequently Asked Questions
What is a foundation model?
A large-scale AI model, typically a large language model, trained on broad, general-purpose data and designed to be subsequently adapted to specific tasks through fine-tuning or prompting, rather than trained from scratch for each new application.
Are foundation models trained specifically on financial data?
Most generative AI tools used in finance today are built on a general-purpose foundation model rather than a model trained specifically and exclusively on financial data, which affects how complete or current its finance-specific knowledge should be assumed to be.
Should a foundation model's finance knowledge be trusted as complete?
No. Because its training data is general-purpose rather than finance-specific, its knowledge of finance conventions, terminology, and edge cases should not be assumed complete or current without independent verification, particularly for anything jurisdiction- or entity-specific.
How is a foundation model typically adapted for a specific finance task?
Most commonly through prompt engineering or by supplying task-specific source data at the point of use, rather than through retraining the model itself specifically for finance.
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.
Large Language Model
A large language model, or LLM, is a machine learning model trained on very large volumes of text to predict and generate coherent, contextually relevant language. LLMs form the basis of most generative AI tools used in finance, drafting, summarisation, and conversational assistants, and their fluency is not itself evidence of factual accuracy, a distinction central to using them reliably in a finance context.
Generative AI
Generative AI is a category of artificial intelligence technique, most commonly a large language model, that produces new language or content, text, summaries, drafted formulas, in response to a prompt. In finance, it is well suited to drafting, summarisation, and narrative tasks, and is distinct from machine learning, which predicts or classifies from structured historical data rather than generating new content.
AI Copilot
An AI copilot is a generative AI assistant, typically built on a large language model, embedded directly within a finance workflow tool, a spreadsheet, an FP&A platform, a reporting system, to support tasks such as drafting, formula assistance, and summarisation through an interactive, conversational interface. A copilot accelerates specific tasks within existing workflow; it does not itself constitute a verification or governance control over the output it produces.