Is AI a mandatory skill area under the Compact?

Many financial services leaders assume that if the government has launched an AI Skills Compact, it must come with legal force behind it.

That assumption is only partly correct.

The Compact is real, it names AI as a priority skill area, and signatory organisations make genuine commitments. But it is not primary legislation, and it does not itself create a statutory duty to train staff in AI.

What it does do is sit alongside, and increasingly reinforce, obligations that already exist under regimes such as Training and Competence and the Senior Managers and Certification Regime.

Understanding that distinction matters, because it changes how firms should prioritise, budget for and govern their AI literacy programmes.

Why this question is being asked now

AI is moving from pilot projects into core operational processes across financial services: credit decisioning, trade surveillance, claims triage, customer service and more.

At the same time, the UK Government's AI Skills Compact has put workforce AI capability firmly on the agenda, encouraging employers to commit publicly to building AI literacy across their organisations.

HR and L&D teams are being asked by boards and audit committees whether this is now "mandatory", often because budget approval and prioritisation depend on the answer.

The honest answer requires separating two things that are frequently conflated: the Compact's own status, and the regulatory obligations that already exist independently of it.

How firms have traditionally handled new skill requirements

Financial services firms are not new to the challenge of embedding an emerging skill area into the workforce.

Training and Competence schemes have long required firms to identify the knowledge and skills staff need to perform their roles competently, and to evidence that training has taken place.

Under SM&CR, senior managers must be able to demonstrate they have taken reasonable steps to ensure the business under their responsibility is run appropriately, which extends to understanding the tools and systems that materially affect customer outcomes.

Role-based training matrices, mandatory CPD hours and periodic competence assessments are the standard mechanisms firms already use to formalise new requirements, whether the trigger is a new product line, a regulatory change or, now, the introduction of AI into decision-making processes.

These existing structures do not need to be reinvented for AI. They need to be extended to cover it.

Where AI-supported learning tools help

Once a firm accepts that AI literacy needs to be embedded through existing training and competence structures, the practical challenge becomes scale.

Different roles need different depths of understanding. An underwriter using an AI-assisted decision engine needs a different level of knowledge to a board member overseeing AI governance, and both differ from a customer service agent using an AI drafting tool.

AI-supported learning platforms can help by adapting content to role, tracking completion and competence evidence, and flagging gaps against a firm's own training matrix. This can significantly reduce the administrative burden of rolling out AI literacy training across a large, varied workforce.

This is a supporting role only. Deciding what staff need to know, to what standard, and how that maps to regulatory obligations remains a governance decision for compliance and senior management, not something a learning platform can determine on a firm's behalf.

What this means in practice for firms

Firms do not need to wait for AI to be declared a formal legal requirement before acting, because the direction of regulatory travel is already clear.

Practical next steps typically include reviewing existing Training and Competence frameworks to check whether AI-relevant knowledge is already implicitly required, mapping roles against the level of AI understanding they genuinely need, and treating the Compact as a strong steer rather than the source of the underlying obligation.

Boards and senior management should also recognise that this is as much a governance and reputational issue as a training one. A firm that can show a considered, evidenced approach to AI literacy is in a materially stronger position than one that has treated it as an optional extra, regardless of how the Compact itself is ultimately formalised.

Example

A London-based retail bank's lending approval team increasingly relies on an AI-assisted credit decision engine.

Internal audit raises a question: are underwriters required to understand how the model reaches its recommendations, and is this covered under the firm's Training and Competence scheme in light of the AI Skills Compact?

The Head of Learning and Development and the Compliance Officer review the position together with the underwriting team.

The firm concludes that while the Compact itself does not legally mandate training, their existing Training and Competence obligations already require staff to understand the tools materially affecting their decisions. They formally incorporate AI literacy into the underwriting competence framework, treating the Compact as a strong steer rather than the source of the obligation.

FAQs

• Is signing the AI Skills Compact legally required for financial services firms?

  No. Signing the Compact is currently voluntary. However, government and regulatory expectations are increasingly pushing firms towards demonstrating workforce AI capability, which makes early engagement a sensible risk management step rather than a legal necessity.

• Does the Compact specify which roles need AI training?

  No. The Compact sets out broad commitments to developing AI skills rather than prescriptive, role-by-role requirements. Firms are expected to map these commitments to their own workforce, based on how AI is actually used across different functions.

• What happens if we ignore AI literacy entirely?

  There is no direct penalty tied to the Compact itself for firms that do not sign or engage with it. However, ignoring AI literacy carries real regulatory, reputational and competitive risk, particularly where existing Training and Competence or SM&CR obligations already require staff to understand the tools that affect customer outcomes.