Don’t let AI happen by accident: Intentionally moving beyond adoption to value creation

I kicked off our recent AI event explaining we’d found about 70 percent of companies in the UK are using AI in at least one business function. Yet 75 percent of those companies aren’t realising the value or the ROI from AI.

What does that mean for business and technology leaders? Our event brought together in-house experts with leading practitioners, investors, academics, and interested parties to explore this question – and look at how to move from adoption to value generation, including the digital and data skills needed to achieve impact.

Achieving impact with intention

After an inspiring keynote from social entrepreneur and computer scientist, Anne-Marie Imafadon, my colleague Derreck van Gelderen, PA’s head of AI strategy, chaired a panel discussion. Alongside Anne-Marie, AI and HR experts from Rentokil, OSB Group and the UK Government’s Department for Science, Innovation and Technology (DSIT) brought unique perspectives to the issue. The wide-ranging discussion threw up five significant areas organisations should address to have an impact.

Recognising the extent of the change

Matt Kemp, AI Portfolio Lead at Rentokil, started by sharing how they’d quickly realised it wasn’t enough to roll out Google’s Gemini to staff. “We needed to communicate, we needed to run proper training sessions – it's the whole change management piece”, he told delegates.

I believe AI is a people and operating model change opportunity, not an IT project – especially now we’ve moved to agentic AI, not just generative. We should design AI to change people’s skills and ways of working – rather than adopting tech for what could be perceived as for the sake of keeping up with others.

It’s not just about technical experts

For your organisation to be intentional about AI, everyone needs a minimum level of competence. Clearly you need some people with deep expertise, but I don’t think the traditional hub and spoke is enough. There’s no one-size-fits-all shopping list of skills, but in a successful organisation everyone should be ‘fluent’ in AI.

While not every leader needs to know how a large language model works, every leader needs to know the impact of AI on strategy, economics, and risk. Understanding risk is vital and could help leaders be less risk-averse, which I’d encourage.

The panel agreed that an organisation-wide growth mindset is vital – because nothing will stay the same. People need to be learning all the time. As Anne-Marie pointed out, we’re talking about AI now, but in years to come quantum will likely become the hot issue and a few years after that, there’ll be something else.

Trust is key

One of the main hurdles to adoption is trust, said Alexis Castillo-Soto, Group Deputy Director and AI lead at DSIT. “It’s a human interaction thing, not a technology thing. People trust AI outside the office, but in the office, they think: no, I'm the expert.” Alexis suggested leadership is all-important. “We’d rolled out AI in [Microsoft] Copilot in one department but discovered the director general didn’t know and had never used it. That was the impetus for us to change our approach. Leaders are the ones people are looking to… they're the ones we need to take on a journey first. And that’s been really useful for us and has led to a flood of ideas coming in.”

Anne-Marie highlighted the power of ‘peer-to-peer’ communication. “I’ve seen things unlock where people compare notes across different departments. If you see something’s working for another team, you think: OK that's working for them and it's been safe and not caused a big kind of ruckus, so it could work in my space.”

Incentives make a difference

One of the challenges when it comes to achieving the desired ROI from AI are fears people have about becoming redundant. Orlagh Hunt, Chief People Officer with OSB group, said: “I’ve heard of cases where people were saving quite small amounts of time, so five minutes or 10 minutes here and there. So, they’d just have a chat or something. It wasn't clearly adding up to enough time to feel it could be purposefully used.”

The solution was to give people a target – for how much time they could save and what they should achieve in that time. “That could be to upskill in AI, say, or look at customer pain points: the things organisations know don't really work for customers, but if solved would fuel growth,” explained Orlagh. Tying AI goals to growth should make people feel more secure in their jobs. And Anne-Marie pointed out that giving people ownership and injecting some competition can also work – when using AI in sales, for example.

Collaboration is where the magic happens

Anne-Marie advocated borrowing from the agile framework to build, measure, and iterate – encouraging people to experiment and make ‘high quality mistakes’. Matt explained how valuable it was to work with colleagues: “We get this technology out to the techs who start to use it every day and then come back and say, yeah, it's cool I like using it, but I wish it could do this. It's that ‘I wish it could do this’ moment that really helps us.”

We’ve long been advocates for having diverse teams, especially when it comes to innovation. Anne-Marie advised getting someone who’s not part of the relevant team, or close to the problem, to consider AI use cases. “Before you approve something for roll-out, you need somebody else in the loop – preferably as random as you can find.” This helps avoid groupthink and can uncover valuable new directions.

It’s about human augmentation, not technical innovation

The discussions provided food for thought on how organisations can close the gap between adoption and value. These five areas – which are all about the human, not about the tech – should help leaders on their journey to creating an intelligent enterprise. A journey taken with intention, not by chance.