Building a Business Case for AI

Ofgem’s Chris Gullick challenges the fantasy of ready-made AI value, arguing that the business case is more of a political, cultural, and organisational project than a document to pitch. 
Building a Business Case for AI

“We can’t do everything for the business. These are your use cases, your benefits. We’re here to help, but the responsibility is shared.” – Chris Gullick, Chief Data and Artificial Intelligence Officer, Ofgem 

 

To gain all the insights of Chris’s keynote presentation and much more, you need to be a DataIQ subscriber. 

 

The reality behind the rhetoric 

Chris has walked into meetings with AI use cases worth £8–14 million in savings, only to be told they fall short of the £50 million vision leaders have in mind. The enthusiasm is real; the grounding is not. For Chris, the issue is not ambition but the absence of focus. His approach is to anchor expectations with concrete examples: where Ofgem sits relative to peers, what comparable organisations have achieved, and where impact is genuinely possible. 

 

“The answer is AI, what was the question?” 

Like many data leaders, Chris faces a constant stream of AI-first solutions: ChatGPT for spreadsheet errors, generic consulting slides, and sweeping claims about AI’s maturity.  

His countermeasure is discipline. In cross-division workshops, discussing solutions is banned and leaders must talk only about problems, outcomes, and pain points. The result was sixty-five use cases and, more importantly, the first real debate among executives about what matters most. 

 

Outcomes beat technology 

Many leaders expect generative AI to dominate the opportunity landscape, but Chris expects the opposite and that most benefits will come from automation, not AI. The Ofgem price-control process illustrates his point. Thousands of pages of regulatory submissions are now being analysed using an AI-supported process, cutting decision-making by months. The significance is not the model, but the industry-wide impact of faster regulatory clarity.  

 

“These aren’t my use cases, they’re yours.” 

Reinforcing business ownership has been one of Chris’s longest battles. Executives happily agree that benefits sit with them, then immediately ask how much he will save. Persistence has been key: repeating the message at board meetings, in one-to-ones, and in every forum available. Over time, non-executive directors have begun to echo it back to the organisation. 

His response is consistent and deliberately uncomfortable: “I’m not going to save any money. You are. I’ll spend money helping you save it.” 

 

Splitting the work 

Chris recognised that leaders want two things: fast progress and big, strategic change. Treating them as one request leads to organisational whiplash, so his answer is a three-stream model:  

  1. Essential AI (quick wins) 
  2. High-Impact AI (deep, focused improvements)  
  3. NextGen Ofgem (the multi-year reinvention) 

It calms the organisation, clears expectations, and protects long-term transformation from short-term impatience. 

 

To access the full article and all of Chris’s insights, plus many more leadership-defining articles, join DataIQ.  

DataIQ 100 Alumni 2026 on stage at the 2026 Congress discussing recognition and leadership.

DataIQ 100 Class of 2026: Turning Recognition into Responsibility

Daniel Susskind at DataIQ Congress 2026

The Future of Work in the Age of AI

Rightmove CEO Johan Svanstrom and CDO Steve Pimblett at Congress 2026.

Turning Data into Business Value: CEO and CDO Insights into Rightmove’s AI Journey

Martin Stockdale on stage at Congress 2026

C-Suite Leading the Hybrid Workforce: Accountability, Decision-Making, and the Rise of AI-Native Organisations

DataIQ C-suite Panel Congress 2026

DataIQ Congress 2026: The End of AI Theatre

From AI Experimentation to Operating Discipline: What Leaders are Learning about Governance

From AI Experimentation to Operating Discipline: What Leaders are Learning about Governance

Beyond AI Experimentation: What Data Leaders are Learning About Scaling AI

Beyond AI Experimentation: What Data Leaders are Learning About Scaling AI

From Insight to Outcomes: Randy Bean opens DataIQ’s Boston Peer Exchange with a Challenge for Leaders

From Insight to Outcomes: Randy Bean opens DataIQ’s Boston Peer Exchange with a Challenge for Leaders

AI generated image of a tropical island set against a moody cloudy backdrop. An AI leader has anchored it to the ground

AI Leadership Today Is About Judgement, Not Just Speed

Quality at Scale: Discover Why Domino’s is DataIQ’s Highest-Scoring Award Winner This Decade

Quality at Scale: Discover Why Domino’s is DataIQ’s Highest-Scoring Award Winner This Decade

We just need 30 seconds of your time

Interested in:

By clicking “Continue” button, you are agreeing to the DataIQ terms of use and privacy statements.