How do we close the AI confidence gap between leaders and staff?

Short answer: UK leaders are far more confident about AI than the people expected to use it, and new Microsoft research published this week puts numbers on that gap. The most reliable way to close it is practical, role-specific AI upskilling for employees that ties training to real tasks, gives people protected time, and is measured on how work changes rather than on attendance.

What the new research found

On 20 October 2026, Microsoft UK released research from Censuswide covering 5,000 UK employees and 1,000 senior decision-makers, with fieldwork run between 2 and 12 October. The headline is a split in outlook:

• Around 90% of business leaders say they have a clear AI strategy.
• 90% of leaders expect AI to have a positive effect on UK society over the next five years, against around 49% of employees.
• Among employees who use AI at work, 35% worry it has reduced their own job security.
• Yet among employees who have used AI, at work or in their personal lives, 71% report a positive effect on their work or career and 69% say it has improved the quality of their work.
• Skills and training made the top three priorities for both groups, chosen by 35% of leaders and 23% of employees.

The same release cites Microsoft's UK AI Diffusion Report, which estimates UK generative AI use rose from 36.6% in the first half of 2025 to 43.7% in Q2 2026.

People are using AI and mostly find it helpful, but many do not trust where it is heading.

Why this matters for L&D, HR and IT

A confidence gap is an adoption problem. Staff who suspect AI threatens their role share less, use unapproved tools quietly and change their habits slowly. That shows up later as unused licences and pilots that never scale.

Two other recent sources point the same way.

First, Workday's October 2026 Global Workforce Report, as reported by Learning News on 9 October, found that advertised demand for basic AI skills such as simple prompting peaked in January 2026 and then fell 25%, while demand for building AI tools and automating workflows rose 51% between September 2025 and July 2026. Workday notes basic skills may simply be assumed now. Either way, a one-off prompting session is no longer enough.

Second, the government-backed employer guide on what works for AI upskilling in the UK (updated 27 July 2026) found that 97% of surveyed organisations already provide some AI training, yet many still report gaps in flexibility (51%) and in practical, contextualised learning (34%). Only 7% had reached the integration stage of adoption and just 1% were scaling.

What actually builds confidence

Confidence comes from doing, not from being told AI is good for you. A few principles hold up well:

1. Start from real tasks. Pick three or four pieces of work each team does every week and train on those. The GOV.UK guide makes the same point: link learning to real tasks and decisions, and be clear about when AI should not be used.
2. Be honest about jobs. Leaders should say plainly what AI is for in your organisation and what it is not for.
3. Give protected time. Asking people to learn AI in the gaps between meetings rarely works. Short sessions of 30 to 90 minutes, which the guide recommends for many settings, are easier to protect than full days.
4. Recognise what people already do. Build on self-taught users rather than starting from zero, and use early adopters as peer champions.
5. Move beyond prompting. Given the shift in skills demand, include workflow design, checking outputs, and simple automation for the roles where it fits.
6. Revisit, do not just deliver. The guide suggests checking in at three and six months.

There may be a compliance angle too. Under Article 4 of the EU AI Act, the AI literacy obligation has applied to providers and deployers of AI systems since 2 February 2025, and the official implementation timeline lists AI literacy among the rules where enforcement started on 2 August 2026. The Act can apply to organisations outside the EU, including UK businesses, so check with legal advisers. Documented, role-based training helps you show reasonable steps.

Practical takeaways by role

L&D leaders: Replace generic AI awareness modules with short, role-based sessions built around live tasks. Our AI upskilling guide sets out a five-step approach and compares formats.

HR leaders: Own the people side of the message. Agree a clear statement on AI and roles with leadership, update your AI use policy, and make sure training reaches frontline and part-time staff. See our page on AI training for HR teams.

IT leaders: Publish a short list of approved tools and plain rules on data, and pair any rollout with training from day one.

Transformation leaders: Agree measures before training starts. Our guide to measuring AI training ROI includes a simple formula and template.

What this means for your team

The gap between leadership optimism and staff confidence is a sign that strategy has run ahead of capability, not a reason to slow down. Close it by treating AI upskilling for employees as ongoing, practical and honest, with leaders using the tools themselves and teams given time to practise on their own work.

If you want a structured way to do that, see how we approach AI upskilling for staff.

Talk to us

If you are planning AI training for your teams and want a second opinion on where to start, book a short chat with Day Seven.