UK organisations have spent heavily on artificial intelligence over the past two years. Licences have been bought, pilots launched and chatbots switched on. Yet for most businesses the promised jump in productivity has not arrived. The latest research points to a reason that has little to do with the technology itself: managers have not been equipped to lead it.

If you manage a team, run a department or sit on a leadership team, this matters to you directly. This article looks at what the 2026 data shows, why the manager is now the deciding factor in whether AI pays off, and the practical skills you can build to close the gap.

What the 2026 research shows

In June 2026 the Chartered Management Institute (CMI) published research based on a poll of more than 1,000 UK managers and interviews with senior business and technology leaders. The findings are striking:

  • 70% of managers reported some productivity improvement from AI, but only 5% described the gains as transformational.
  • 26% had seen no productivity gains at all.
  • 68% of organisations were still testing AI or running early-stage pilots.
  • Only 12% of managers felt “very confident” managing teams that use AI, falling to 10% for more advanced agentic AI systems.

As the chair of the CMI’s AI advisory council put it, AI adoption is “a management challenge”, not only a technical one. You can read a summary of the findings in Workplace Journal’s report on the CMI research.

Research published by Kaplan UK in October 2026 tells a similar story from a different angle. In a survey of 2,200 people, including 1,000 employees and 600 company directors and C-suite leaders, 54% of C-suite respondents said their organisation’s investment had favoured technology over employee capability. Around a third of employees said they had been asked to work with AI without sufficient training, and 37% of senior leaders said an AI tool or deployment had failed to deliver its expected return. The full figures are in IT Brief’s coverage of the Kaplan study.

The pattern is consistent: the tools are in place, but the people who decide how work gets done have not been given the knowledge or confidence to redesign that work around them.

Why managers, not IT, decide whether AI pays off

It is tempting to treat AI as a technology project owned by the IT team. In practice, the value of AI is created (or lost) in everyday management decisions:

  • Which tasks change. Someone has to decide which parts of a workflow AI should handle, which need human judgement, and how quality will be checked.
  • How people are supported. Teams adopt new tools when their manager uses them, explains the purpose and makes it safe to experiment.
  • How risk is handled. Data privacy, accuracy and bias are business risks. Managers need to know when an output can be trusted and when it cannot.
  • How success is measured. Without clear measures, pilots drift on without ever proving their value, which helps explain why so many organisations remain stuck at the testing stage.

None of these decisions require coding skills. They require leaders who understand what AI can and cannot do, and who can translate that into strategy, process and people management.

Five AI leadership skills to build in 2026

1. Understanding AI well enough to ask the right questions

You do not need to know how a model is trained, but you do need to understand the difference between predictive AI, generative AI and automation, where each fits, and their typical failure points. This is what allows you to challenge vendor claims and set realistic expectations.

2. Linking AI to business strategy

The most effective leaders start with a business problem, not a tool. Which processes are slow, costly or error-prone? Where would better information speed up decisions? Prioritising a small number of high-value use cases, and estimating their return, keeps AI investment focused.

3. Redesigning work and roles

AI changes the shape of jobs. Leaders need to map tasks, decide what is automated or assisted, and rethink roles, training and workloads accordingly. Done well, this frees people for higher-value work; done badly, it creates confusion and resistance.

4. Governance, ethics and risk

Clear rules on data use, confidentiality, human review and accountability protect the organisation and build trust with staff and customers. Managers should know their organisation’s AI policy and be able to apply it to real situations.

5. Leading change and building confidence

The research shows a confidence gap at every level. Leaders who use AI themselves, share what works and give teams time to learn create the conditions for adoption. This is classic change leadership applied to a new context.

A practical 30-day plan for managers

  1. Week 1: Audit. List your team’s main tasks and note where AI tools are already being used, officially or not.
  2. Week 2: Choose one use case. Pick a single, measurable task where AI could save time or improve quality, and agree how you will measure success.
  3. Week 3: Set guardrails. Confirm which data can be used, who reviews outputs and how errors are reported.
  4. Week 4: Review and share. Measure the result against your baseline, gather feedback from the team and decide whether to scale, adjust or stop.

Repeating this cycle builds evidence, confidence and momentum, which is exactly what the 2026 data says most organisations are missing.

Build your AI leadership skills with London School of Business

London School of Business offers executive education designed for leaders who need to understand and apply AI without a technical background. Our AI for Business Leaders and Executives professional certificate covers AI strategy, investment and return, and the practical use of generative AI for productivity and decision-making. It is available online or as a one-day course on campus in London.

For a deeper programme, the six-month Diploma in AI for Leaders and Executives covers ten modules, from AI strategy and data governance to risk management, workforce transformation and change management, finishing with a capstone AI strategy roadmap for your own organisation. It can be studied online, in a live virtual class or on campus.

Explore our full range of executive education courses, or request more information and our team will help you choose the right option for you or your organisation.