The right model for digital and AI capabilities – James Craig speaks at the Private Equity International Operating Partners Forum

At the recent Private Equity International Operating Partners Forum in London, our Portfolio Growth Director James Craig joined a panel of operating leaders to discuss the challenges of AI adoption and integration in PE-backed companies. The conversation kept returning to two things: getting change management right and measuring performance.

Why AI is changing who does the work, not just how it gets done

In the opening, the panel discussed how previous waves of digital transformation were about changing which systems a business ran on, but AI is fundamentally different: it changes who does the cognitive work, not just which system processes it. That means the way people work and the processes they follow have to change with it, this is not a system swap, it is an operating model shift.

It is also happening faster than any previous wave. Adoption is accelerating across PE portfolios, but many businesses are still stuck in pilots.

From IT initiative to operating model priority

The common theme everyone agreed on was that businesses are still treating AI as an IT project, delegated to the technology team, with a roadmap requested and a working group formed. Whereas it needs to sit with the board as an operational and commercial priority, owned by senior operational leaders, measured through productivity and business outcomes, and reflected on the P&L rather than confined to the IT budget.

Change management and measurement matter

Drawing on his background as a consultant, James emphasised that AI initiatives only succeed when change management matches the pace of change. That sounds obvious, but in practice it is where most come unstuck. It cannot be a side-of-desk task. Someone has to own it and be accountable for it, with the same rigour as the technical delivery – a named internal champion, usage tracked in one-to-ones, and new ways of working embedded into existing processes rather than bolted on alongside them.

Measurement, James concluded, is the other half of the same problem. Boards should be looking past adoption metrics and towards business outcomes: how AI is improving cost, speed, quality or productivity. The single most practical step is to put an AI metric into every portfolio company’s monthly reporting pack. Ultimately what gets measured gets delivered.

The direction of travel is already visible in GCP’s own portfolio as we adopt AI capabilities. AI initiatives are no longer a discretionary investment; they are becoming a structural capability of any PE-backed business. The overall sentiment was that in three years’ time we won’t distinguish between ‘AI-enabled businesses’ and everyone else. There will simply be businesses that have successfully embedded AI capabilities into how they operate – and those that have not.