Closing the AI readiness gap starts with skills

The BBC reports that 46% of London’s workforce, around 2.4 million people, are in roles where AI could automate part of their work. At the same time, employers are finding it harder to access the skills they need. A BusinessLDN survey found that only half of employers believe their existing workforce has the required skills, down from 63% a year earlier. 

This gap signals a fundamental change in how jobs are done, particularly as AI takes on more routine, administrative, and data-heavy tasks. The critical question is whether workers will have the skills and confidence to benefit from that change. 

Access doesn’t equal readiness 

QA’s analysis shows that AI readiness varies sharply by sector. Technology and IT lead on adoption, while financial services has seen the fastest growth since 2022, while professional services organizations are using generative AI for drafting, research, and knowledge work. 

In manufacturing, 41% of AI deployments are being driven by labor and skills shortages, rather than efficiency alone. This suggests employers are increasingly using AI to address capability gaps and help existing teams do more. 

Healthcare faces a different challenge. Although its AI market is growing, research among frontline clinicians found that most had never used AI in their work, with fear of clinical error the leading barrier. This is a trust gap as much as a technology gap. 

These findings show why measuring access or adoption alone can create a misleading picture. An organization might have rolled out AI tools without giving employees the practical skills, critical judgment, or confidence to use them well. 

Start with an honest view of capability 

There’s no single AI transition. Different sectors, organizations, and occupations are moving at different speeds, so employers need to understand their own starting point. 

That means identifying which roles and tasks are changing, where AI could add value, and which capabilities employees already have. It also means finding the gaps. Do people understand how to assess AI-generated outputs? Can they recognize when a tool shouldn’t be used? Are managers equipped to support responsible experimentation? 

London’s proposed early-warning system could help employers and policymakers anticipate changes before they become more serious workforce problems. But the data should track more than roles exposed to automation. It should also identify emerging skills, changing tasks, and new career pathways. 

Human skills are part of AI capability 

As AI takes on more repeatable work, skills such as critical thinking, communication, problem-solving, and judgment will become even more valuable. 

These human capabilities aren’t separate from AI skills. An employee may know how to produce an answer using an AI tool, but they’ll still need to evaluate its accuracy, identify what’s missing, and adapt the output for the right audience. They’ll also need to understand the risks and decide whether using AI is appropriate in the first place. 

Give every role the right AI skills 

Workforce development, therefore, can’t be limited to short demonstrations of new tools. Employees need practical, role-relevant opportunities to build their skills. Managers also need training so they can create safe environments for experimentation, recognize good use, and intervene when safeguards are needed. 

The aim isn’t to turn every employee into an AI specialist. It’s to give people the right level of AI literacy, practical capability, and human judgment for their role. 

Workers should help shape what comes next 

London’s plan recognizes that workers must be involved in decisions about how AI changes their jobs. That involvement will be essential for building trust and making sure new systems improve work in practice, not just on paper. 

Employees understand where processes create friction, which tasks carry the greatest risks, and where human expertise adds the most value. Giving them a role in redesigning work can uncover better uses for AI while reducing the sense that change is simply happening to them. 

This is particularly important for entry-level roles. If AI takes on the routine tasks through which people traditionally gain experience, employers will need to create new ways for early-career workers to learn and progress. Otherwise, today’s productivity gains could weaken tomorrow’s talent pipeline. 

Measure capability, not just rollout 

The real measure of AI progress isn’t how many people have access to a tool. It’s whether they can use it safely, confidently, and productively. 

Organizations should measure capability, confidence, behavior, and outcomes alongside adoption. As Jo Bishenden, Chief Learning Officer at QA, explains: “Adopting AI tools is the easy part. Building the confidence and skills to use them well is where the real magic happens. But it doesn’t happen automatically.” 

London may be at the sharp end of AI-driven change, but the challenge extends across the UK. The organizations that benefit most won’t simply be those that adopt AI first. They’ll be the ones that understand their skills gaps, involve their people, and invest deliberately in workforce capability.

Get in touch to discuss AI skilling today