L&D's Next Move

Last week, we highlighted why AI literacy is no longer optional. Powering smarter decisions, stronger governance and organisation-wide readiness, AI literacy is now a strategic imperative across every function and at every level throughout financial services firms.

Here, we turn our attention to what this means in practice, both strategically and operationally, and outline the critical steps learning and development teams must take to equip financial services firms for confident, compliant and competitive AI adoption.

Future‑proofing teams and skills: what L&D must do next

AI is already rewiring BAU within financial services, from client onboarding, surveillance and transaction monitoring to regulatory reporting and boardroom conversations. The firms that win will not be those that simply buy tools, but those whose people can competently question, govern and operationalise them. For L&D, that means moving fast, compounding one‑off courses with continuous, role‑based capability building that matches the speed of change.

Why future‑proofing matters now for financial services firms of all sizes

Boards are investing, employees are experimenting and regulators are watching. Yet capability is patchy. Shadow AI is rife because many staff find unsanctioned tools that make their work easier and use them anyway. Bans fail. Education, policy and monitoring succeed. At the same time, leaders face opinion‑heavy, evidence‑light debates because directors and senior managers lack a shared mental model of how AI really works and where it fails.

Work itself is changing shape. Entry‑level and some middle management roles are already being squeezed as automation takes over routine coding, data prep and MI production. Indeed, there is already line of sight to organisational structure shifts from pyramid to diamond, with fewer juniors, a bulge of higher judgement roles and a leaner top.

The L&D remit: capability at the speed of need

L&D programmes cannot centre on tools alone. Provision must help every function learn to frame problems, interrogate data, validate models, monitor drift and evidence decisions. Learning needs to be purpose-built, strategy-specific and assessed for judgement, not rote recall. Think adaptive pathways, red‑team simulations, regulator‑style interviews and model documentation labs.

Key moves:

• Map roles to proficiency tiers (Board, ExCo, Risk & Compliance, Model Owners, Frontline). Not everyone needs the same depth, but everyone needs enough to be accountable.

• Teach with real use cases across onboarding, monitoring and reporting so staff can see how controls, explainability and data lineage show up in practice.

• Normalise safe experimentation. Provide sanctioned tools, logging and guidance to displace shadow AI rather than fight it.

• Embed sustainability and cost literacy. AI is energy hungry. Leaders and builders must understand the operational and ESG impact of scaling models.

A practical skills blueprint for regulated firms

• AI foundations for operators: What different model classes can and cannot do, common failure modes, prompt hygiene, verification habits.
• Risk, ethics and regulation for practitioners: EU AI Act risk tiers, ISO 42001, FCA expectations, bias detection, data minimisation, model cards, audit trails.
• Lifecycle governance for builders and buyers: Inventory, approval gates, RACI, ‘explainability’ thresholds, monitoring metrics, retirement plans.
• Productivity with control for the many: Copilot‑style assistants, MI automation, policy‑aligned usage, documented checks, escalation points.
• Leadership for an AI‑accelerated org: Decision‑making with real‑time MI, working with narrow intelligences as “AI board advisers”, communicating transparently with the whole workforce.

90 days to visible progress

0–30 Days

• Conduct a C-suite and Board briefing to define enterprise-wide AI ambition, risk appetite, and accountability pathways, with reference to FCA Principles for Businesses, ISO 42001 and Consumer Duty obligations
• Establish a real-time inventory of all AI technologies in use across the firm, including embedded AI in third-party platforms
• Map active and emerging AI use cases by business unit and prioritise based on impact and regulatory risk
• Publish a concise, plain-English Responsible AI policy, aligned to ISO 42001 and enhanced with sector-specific guardrails
• Nominate second-line leads (Risk, Compliance, Legal) to participate in governance design, documentation reviews and risk classification

31–60 Days

• Launch role-based learning paths for priority cohorts:
• Frontline: Prompt design, use case ownership, policy limits
• Second-line: AI-specific risk reviews, oversight protocols
• Model owners/Developers: Monitoring, validation, documentation
• Introduce model documentation templates, including data lineage and consent pathways, validation and performance metrics and explainability and model update protocols
• Deploy sanctioned AI workbenches or sandboxes to displace shadow AI tools, control access and identity, log usage and monitor prompt content
• Run a second-line-led review of AI artefacts to check documentation, materiality thresholds and regulatory alignment
• Establish baseline model governance routines: versioning, approvals, audit logging

61–90 Days

• Run a red‑team exercise against a real workflow for example, market abuse surveillance or regulatory reporting
• Capture red team learnings into policy and training updates, with specific triggers for escalation
• Produce the first AI MI pack to the Board: adoption rates by business unit, risk indicators, override rates, bias metrics, energy cost
• Calibrate judgement-based assessments to test judgement under e.g. regulatory scrutiny, not tool familiarity
• Introduce continuous assurance mechanisms including prompt audit trails, model drift detection, real-time alerts for threshold breaches

This 90-day sprint establishes foundational AI governance and capability maturity. It transitions into a quarterly AI governance rhythm, with repeatable assurance reviews, thematic risk reporting, and annual benchmarking of organisational AI oversight.

Pitfalls to avoid

• Training people to use a tool, rather than thinking with and challenging it
• Delegating AI upskilling to Compliance alone
• Underestimating culture: if experimentation is punished, shadow AI will thrive
• Creating an ethics or governance forum with no power or throughput, which teams learn to route around

Final thought

Future‑proofing is not simply an exercise in course-completion. It is a continuous, governed, integrated capability system that lets people learn, test and evidence safe AI use in real work. L&D is pivotal to making that real. Start small, move fast, measure everything and keep closing the loop between training, policy and practice.

Over the coming weeks we will continue to unpack the remaining change areas:

• Ethics & Responsible Use
• AI Governance

Sign up now to receive the latest insights to your inbox [link]. You can also follow us on LinkedIn [link] to stay informed and connected as the conversation develops.

Want to equip your workforce to adopt AI safely and confidently, from Board to first-line? Contact us to discover how ZISHI’s AI training programmes, coaching and consultancy – designed specifically for financial services – can accelerate your organisation’s AI readiness, integration and onboarding goals info@thezishi.com.