Building an AI-ready workforce

For all the breathless headlines about artificial intelligence, the inconvenient truth for leaders is this: enterprise AI is racing ahead, faster than people, processes and governance can follow. 

 

Almost every large UK organisation is investing in AI, yet very few have translated pilots into scaled, value-creating workflows. According to McKinsey, almost all companies reported investing in the technology, but only 1% considered themselves “AI mature” (where AI is fully embedded and delivering material outcomes). The research points bluntly to leadership and organisational readiness - not employee appetite or core technology - as the binding constraint.

 

The UK’s own AI sector is also booming, but the public evidence is equally clear that skills and adoption capacity lag behind the technological frontier.

 

 

Why AI strategies fail without workforce readiness

It’s tempting to view the AI gap as a tooling problem, solvable with more models, more compute and more vendors. But data suggests otherwise: 

• Technology is ready; operating models aren’t: McKinsey’s 2025 State of AI finds broad use case experimentation, but two thirds of organisations haven’t begun scaling across the enterprise, because workflows, incentives and guardrails haven’t been rewired to absorb AI.
• Employees are using AI - often quietly - without institutional support: Studies show widespread ‘BYOAI’ (bring your own AI) and shadow use. Microsoft’s Work Trend Index reported that 75% of knowledge workers use generative AI at work, while many leaders admit their organisations lack a plan. On top of this, another report found 8 in 10 employees use AI regularly, yet over a third of UK workers report no training, and most don’t discuss their usage due to unclear policies.
• Leadership and literacy gaps erode value: EY’s 2025 Work Reimagined survey concludes firms are missing up to 40% of potential AI productivity gains because talent strategies, culture and training are underpowered. And only 12% of employees say training is sufficient to unlock AI’s benefits. 

These findings resonate with OECD’s 2025 brief: training supply, especially for general AI literacy, is falling short of demand across advanced economies. In other words, technology is no longer the main bottleneck; workforce readiness is.

 

 

What leading organisations are doing differently

The most effective organisations treat AI reskilling as a change journey, not a learning event. Five patterns stand out here: 

1. They start with business outcomes, then back solve to skills and roles‑solve to skills and roles. Rather than ‘AI for everyone’ training, leaders prioritise a small number of high-value workflows, define the skills to deliver those outcomes and sequence learning accordingly. This ‘goals before roles’ principle prevents vanity metrics and accelerates time-to-value.
2. They redesign workflows and incentives…then add tools. High performers incorporate AI directly into standard operating procedures and adjust KPIs to reward adoption, quality and safety, not just speed. Workflow redesign is a hallmark of organisations reporting enterprise-level impact.
3. They invest in managers as enablement nodes. The ‘silent use’ of AI persists when managers can’t coach, assess quality or approve use. Organisations tackling this head on provide manager playbooks: what ‘good’ looks like, how to review AI outputs, how to discuss provenance and when to escalate to experts, addressing the cultural and policy ambiguity that UK workers report.  
4. They professionalise governance early. Leaders codify allowed/blocked use cases, data handling rules and human-in-the-loop thresholds, and they equip employees to spot model failure modes. Such progress to clear guardrails enables safe experimentation at scale.
5. They leverage national frameworks and public programmes. UK organisations aren’t starting from scratch. The AI foundation skills benchmark and Skills England’s employer tools (AI Skills Framework, Adoption Pathway, Employer Checklist) translate policy into practice, and offer a useful scaffold for role maps, curricula and measurement. 

 

Ensuring ROI on reskilling

Ensuring a meaningful return on investment from AI reskilling requires organisations to anchor their learning strategy to economic outcomes and risk posture, rather than simply counting hours of training delivered.

 

The most effective programmes begin by establishing clear ROI guardrails. This means defining, for every AI-enabled use case, both the minimum level of value it must generate - such as cost savings per case or measurable improvements in quality - and the maximum risk the organisation is willing to tolerate - whether that relates to hallucination rates, data handling errors or other forms of exposure.

 

Organisations also need to monitor adoption where it actually happens: in the workflow, not the classroom. Instead of relying on completion rates or satisfaction scores, leaders should track how employees are using AI tools in their day-to-day roles by measuring usage patterns, exception handling, approval flows and the quality of outcomes. This operational visibility allows them to link skill uptake directly to tangible business results, such as reduced cycle times, increased revenue contribution or fewer risk incidents.

 

Finally, ensuring ROI demands a different approach to budgeting is also key. Instead of treating AI-related training as a fixed cost or tying investment to software licences, organisations should allocate resources based on outcomes achieved, reinvesting the savings generated by early wins into the next wave of capability building. This creates a self-sustaining transformation cycle in which learning continuously funds further innovation.

 

Organisations that frame AI adoption around growth and innovation objectives - supported by investment models that reward realised value - consistently capture more impact than those focused only on efficiency. The UK now has the policy framework to accelerate readiness: the AI foundation skills programme open to every adult, the ‘10 million by 2030’ upskilling target, and a sector-level view of future digital and AI talent needs. The question is execution.

 

AI won’t replace people; but people and organisations that learn to work with AI, systematically, responsibly and at scale, will outpace those that don’t. Businesses should treat reskilling as its primary AI programme. The technology is ready. Your workforce can be too. 

 

---

 

Professor Serkan Ceylan is Dean of Faculty of Business and Innovation, Arden University

 

Main image courtesy of iStockPhoto.com and MicroStockHub