Codebreakers and change-makers

When business-critical software can be built in hours, what happens to organisations designed to make decisions in months? And if AI expands our capacity tenfold or more, how do we decide what is worth doing?

These questions ran through the Enterprise AI Summit, hosted by IT Revolution in partnership with Newton, which brought more than 170 people from defence, healthcare, policing, infrastructure, government and retail to Bletchley Park on 2–3 November. 

Bletchley Park brought people together to crack a complexity no single mind could solve alone and our first expert speaker of the day, Jacqueline de Rojas CBE, Chair of the Bletchley Park Trust and President of Digital Leaders, set the tone drawing comparisons with the 9,000 people who came to Bletchley during the Second World War to break encrypted enemy communications at speed and scale. The aim was a different kind of AI event, with speakers chosen for what they have delivered rather than their title, and with every talk covering what was tried, what worked, what failed and what it cost. 

The atmosphere was lively from the start. Speakers were open about what had failed and what it had cost and each closed by naming what they wanted help with from the room. They stayed after their talks and joined the conversations that followed. Delegates from very different sectors used a shared challenge to ask each other for advice and insight they would not normally find in their own networks.  

• Ambition beyond productivity 

Speakers repeatedly challenged the idea that the prize from AI is simply doing today’s work faster or with fewer people. Gene Kim encouraged delegates to look beyond productivity gains of around 20 per cent and ask a bigger question, what could organisations use 10X gains to achieve that wasn’t possible before? 

• Redesign the system, not just the process 

As AI accelerates individual activities, the constraint can simply move somewhere else. Gus Paul made the point that speed in one part of a business can create a bottleneck in another and Matthew Skelton talked about reshaping teams, workflows and incentives. 

The bigger opportunity is not to apply AI to individual processes in isolation, but to step back and look across the organisation as a whole, how its operating model works, where silos exist, how decisions are made, how governance works and how activity flows through the system. That creates the chance to redesign the system around what AI now makes possible. It is when AI becomes part of strategic transformation, rather than being treated as the technology answer, that organisations can start to unlock the biggest impact. 

• AI is changing the job of leadership 

As AI makes more information available, surfaces signals earlier and gives people across an organisation a greater ability to act, it changes what is required of leaders. Dame Emily Lawson’s discussion with Mallory Durran explored this shift, leaders don’t need to have all the answers themselves, but they do need to create the conditions for better and faster decisions across the organisation. That means being clear about outcomes, putting decision-making closer to the people with the right knowledge, and being willing to learn and adapt as the technology develops. 

As Ben Grinnell, Newton’s Chief AI and Capabilities Officer, put it on day one: “We need to democratise decision-making to the place where the people know best. Move decisions to data not data to decisions”. 

• AI adoption is a human challenge as much as a technical one 

Again and again, speakers described organisations or teams with access to similar technology achieving very different results. Confidence, behaviours, incentives and people’s willingness to question established ways of working all matter and AI is loosening the boundaries between roles, with developers working across specialisms and non-technical colleagues contributing directly to products. 

At the same time, examples from high-stakes environments showed that experimentation and speed don’t need to come at the expense of safety, security or assurance. Faster learning loops are critical for both and leaders increasingly have to create the conditions for both. 

There was also a distinctly UK dimension. Regulation, security and safety sat alongside a strong focus on outcomes for people. Justin Geldof, Newton’s Technology Director, who has attended US summits, reflected on how UK regulation, governance and risk shape many of the same questions differently. The interesting point was not that the UK conversation was more cautious, but that ambition and responsible adoption were being considered together. 

In the breakout groups, delegates asked how to tell whether AI is supporting their judgement or replacing it, where the next senior engineers will come from and how to separate business value from a rise in output. Gene Kim closed by saying the summit had shown what is possible and that the question now is how much ambition and imagination leaders can bring to the outcomes that really matter. 

The conversations across the two days reinforced something we see in our own work. For 25 years, Newton has been about getting under the skin of complex systems, understanding how they really work and finding a way through to impact that lasts. AI doesn’t change that purpose. It gives us new ways to do it and the potential to take on complexity that was out of reach before. 

We heard repeatedly that the biggest opportunities won’t come from treating AI as the answer in itself, or from applying it to individual processes in isolation. They come from combining what AI makes possible with a deep understanding of the organisation around it and redesigning how that organisation works to achieve much bigger outcomes. That is where we are focused, bringing leading AI expertise to the organisational complexity our clients actually face and turning it into meaningful impact.