Brooke Johnson, Chief Legal Counsel, SVP of HR and Security, Ivanti, says despite billions in UK investment, most businesses are failing to realise meaningful returns from AI due to poor governance, unclear ownership and lack of visibility.
The Chancellor’s £2.5 billion package for AI and quantum, set out alongside her March Mais lecture, makes the government’s position clear.
The UK wants the fastest AI adoption in the G7, and British businesses are being actively encouraged to spend. What happens after the money clears is where enterprise IT leaders will earn their keep, because committing budget is a simpler problem than delivering outcomes.
UK businesses have already put between £22 and £30 billion into generative AI, according to MIT’s State of AI in Business 2025 research, and only 5% report meaningful returns. The disappointing results so far have less to do with the technology than with how it has been introduced, owned and measured inside organisations. It’s clear to me that the next wave of investment will produce the same numbers unless that changes.
Visibility
According to Ivanti’s Technology at Work Report, 57% of IT professionals believe that leadership lacks clear visibility into IT efficiency and its underlying drivers. This is just one example of why introducing AI into environments that are already opaque doesn’t just add complexity – it amplifies existing challenges.
As tools proliferate, ownership becomes muddled, budgets are harder to track and security and compliance risks escalate rapidly, outpacing any single department’s ability to manage them.
Ownership
Who, specifically, owns AI governance at a leadership level? That’s an underrated but really useful question we should ask ourselves, as well as our peers in other organisations. The answer is often a little awkward, because responsibility tends to be shared across IT, security, legal and individual business units in a way that means no single function has the authority or reach to make decisions that stick.
Governance does not have to live in any one department, but it does need one clear line of accountability with visibility across the business. Without that, policies exist as documents and are rarely practised consistently.
Shadow AI
The same research found that 46% of office workers are using AI tools their employer has not provided, and 32% of AI users keep that activity hidden from management. I’d argue that treating this as an employee discipline issue misreads the problem. When approved tools are slow to arrive or unclear in what they can and cannot be used for, people reach for whatever helps them get the job done.
Don’t reach for tighter prohibition as the fix. What truly needs to happen is faster, clearer sanctioned alternatives paired with practical guidance on what is allowed in which context.
The UK has a genuine opportunity within the current investment moment. The only way to take advantage of it is to pair spending with structure. That 5% number should be far higher. The only reason it’s not zero percent is because those companies have done the operational work of assigning ownership, building visibility and integrating AI into how the business runs.
Everyone else has the same opportunity, provided the next round of budget goes somewhere it can be governed and measured.
