What Is the Business Case for Investing in AI Literacy?

Every financial services firm now faces the same question: who in the organisation actually understands the AI tools already being used in trading, underwriting, claims and compliance?

AI adoption is moving faster than workforce understanding. Traders, underwriters and operations staff are using AI-assisted tools daily, often without a clear grasp of what those tools can and cannot be trusted to do.

Leaders are increasingly asked to justify why AI literacy deserves dedicated budget rather than being folded into general digital skills training.

This article sets out the operational and regulatory case for treating AI literacy as a funded investment, not a discretionary extra.

Operational Context

Across London's financial markets, AI tools have moved from pilot projects to everyday use. Pricing engines assist trading desks, natural language tools draft underwriting notes, and machine learning models flag claims for review.

This adoption has outpaced the workforce's understanding of how these tools actually work. Staff frequently know how to operate an AI tool without understanding its confidence limits, failure modes or the judgement still required around its output.

The resulting gap is not theoretical. It shows up as traders treating AI-assisted pricing as authoritative rather than advisory, underwriters accepting AI-drafted summaries without checking source data, and compliance teams unable to explain how an AI-influenced decision was reached when a regulator asks.

Each of these is creates operational exposure. And as the gap widens between AI adoption and AI understanding, the larger the exposure becomes.

Traditional Approaches

Firms have historically justified workforce training investment through three familiar routes: compliance mandates, skills gap analyses and competency frameworks.

Compliance-driven training works well when a regulator specifies exactly what must be taught and to whom. Skills gap analyses work well when the required skill is stable and well understood, such as a new product line or regulatory change. Competency frameworks work well when roles and responsibilities are clearly defined and slow to change.

AI literacy does not fit neatly into any of these models. Regulatory expectations are still forming rather than fully specified. The required skill set is not stable, since AI tools and their capabilities change frequently. Roles are shifting as AI reshapes tasks within trading, underwriting and customer service.

Applied unchanged, traditional training justifications tend to underfund AI literacy. It gets treated as a generic digital skills top-up, scheduled once and rarely revisited, when the underlying risk is continuous and evolving.

Where AI Helps

AI literacy is a distinct capability from general digital literacy, and this distinction matters for the business case itself.

General digital literacy covers using software competently. AI literacy covers something different: understanding when an AI output should be trusted, when it should be checked, and when it should be overridden. It includes recognising the limits of a model's confidence, understanding where bias or data quality issues might distort an output, and knowing what governance obligations apply when AI contributes to a decision.

The same AI systems that create this literacy requirement also generate data that can strengthen the business case for addressing it. Usage patterns, error rates and near-miss incidents involving AI tools are measurable. A firm that tracks how often staff override AI recommendations, how often AI-assisted decisions require correction, or how frequently near-misses occur can quantify the cost of low literacy directly, rather than relying on general assertions about risk.

This turns the business case from a qualitative argument into a quantitative one. Leadership can compare the cost of a structured literacy programme against the measurable cost of AI-related errors, exceptions and rework already occurring across the business.

Operational Considerations

A credible business case for AI literacy investment should reflect a few practical realities.

• Investment should be phased and role-specific. Trading and underwriting teams making AI-assisted decisions need a different depth of literacy than back-office staff using AI for routine document processing.
• The case should tie directly to existing risk registers and regulatory obligations, rather than sitting as a standalone HR initiative that competes with other training budgets on soft criteria alone.
• Leadership sponsorship matters. Programmes that start as a reaction to a single incident need a named executive owner to survive beyond the initial pilot.
• Outcomes should be measurable and reported against existing performance and risk metrics, so continued funding is a data-led decision rather than a renewed argument each year.

Taken together, these considerations turn AI literacy from a one-off training expense into a monitored operational investment, comparable to how firms already manage other forms of operational risk.

Example

A London-based commodity trading desk begins using an AI-assisted pricing tool to support settlement decisions.

Traders with limited understanding of the tool's confidence limits begin treating its outputs as authoritative rather than advisory. This leads to a settlement discrepancy that requires manual unwinding and internal audit review.

Operations leadership uses the incident to build a business case for a structured AI literacy programme, quantifying the cost of the discrepancy and comparing it to the cost of targeted training for the desk.

The firm secures budget for a phased programme prioritising trading and settlement teams first, using the incident cost as the baseline justification. Subsequent reduction in AI-related settlement discrepancies becomes the measure of return on that investment.

FAQs

• Isn't AI literacy just another name for general digital training?

  No. General digital literacy covers using software competently. AI literacy covers a different set of risks specific to AI, such as over-reliance on outputs, misunderstanding confidence levels, and the regulatory obligations that apply when AI contributes to a decision. Generic digital training does not address these risks.

• How do we quantify the return on investment for AI literacy?

  ROI can be framed through reduced error rates, faster and safer adoption of AI tools, lower compliance exposure, and improved staff retention. Benchmarking these outcomes against the cost of incidents caused by low AI literacy, such as settlement discrepancies or compliance breaches, gives leadership a defensible, measurable comparison.

• Does the Skills Compact make AI literacy investment mandatory?

  The Skills Compact signals a clear government and regulatory expectation that firms invest in workforce AI capability. It is not a training preference. Firms that invest early get ahead of compliance requirements rather than reacting to them once expectations are formalised further.