The Quiet Revolution

By Gary Joyce, Managing Partner at Genesis

Something shifted for us in Genesis over the last eighteen months and we’ve been trying to find the right way to describe it to clients and keep landing on the same word: structural. This is like the industrial revolution at breakneck speed!

The early wave of AI was about productivity – at Genesis, we used it to do familiar things faster. What's emerging now is different in nature. The systems coming through now take a brief and run with it - working through problems, coordinating tasks, acting across whole workflows with minimal human input. A single person can now oversee a constellation of specialised systems working in parallel, drafting analysis, stress-testing assumptions, monitoring regulatory change, preparing board-ready briefings - while focusing on the judgements that require a human in the room. For organisations of any size, in any sector, most people haven't fully reckoned with what that means for how they're structured and what their value proposition is.

We know this not just because we advise organisations going through it but because Genesis is going through it too.

We're a small professional services firm and we know that if we don't adopt agentic AI seriously, we won't be competitive. A meaningful portion of what professional services firms have traditionally charged for is now replicable at a fraction of the cost and time. We're not exempt from that, so when we sit with a leadership team and ask hard questions about strategy, positioning or organisational design in the era of AI, we're asking ourselves the same questions. That, I think, makes us more useful and more real as a conversation partner.

The question we keep coming back to, for our clients and ourselves, isn't "how do we adopt AI?" It's the one before that: ‘if we were starting this organisation from scratch today, knowing what we know now, what would we do differently or better?’

It's a deliberately uncomfortable question. Most organisations approach AI as something to layer onto existing structures, for example, automation applied to current workflows. But a lot of those workflows exist because of constraints that no longer apply. Why does it take multiple reviews to produce a board recommendation? Why is the organisation shaped the way it is? The blank-page question forces leadership teams to separate what needs to stay from what's the residue of how things happened to evolve.  (I could say the same thing about the old 9-5 workday and the 5-day working week!).

We've been running this as a facilitated exercise with some of our clients and what's consistent is how quickly the conversation stops being about technology. It turns almost immediately into something more searching: about purpose, about which parts of the organisation exist because they genuinely need to, and which exist because they always have.

We ask teams to think about it in three ways. The customer or user experience they'd design if legacy systems weren't a factor. The team they'd build if they were recruiting from Monday. And the decision rights they'd establish if no one had ever held those roles before. The gap between that picture about the dream organisation and where they are today tends to be more clarifying than any technology roadmap.

Which brings me to governance, because this is where boards need to be ahead of the curve rather than catching up. Most leadership teams we speak with have some version of an AI strategy on paper. The harder question is whether they have governance fit for a world where systems are acting autonomously rather than just generating drafts for humans to review.

An autonomous agent making decisions across procurement, hiring or stakeholder communication raises questions of accountability that aren’t easily resolved. We've found it useful to get boards to answer five questions: Who is responsible when an agent acts? Can we reconstruct, in plain language, why a decision was made? Which judgements are explicitly reserved for humans, and is that written down? What happens when an agent fails or produces something wrong? And are we building the human capability to oversee what we've deployed? Most boards can answer two or three of those with confidence, but many cannot. The unanswered questions are a reasonable place to start.

The deeper point and the one we keep coming back to most often is that as AI absorbs more of the routine cognitive work, what remains distinctly human becomes more valuable. Judgement, ethical reasoning, creativity, empathy, the ability to read a room and know what isn't being said - these are the irreducible parts of leadership, in any sector. The organisations that navigate this well will be the ones that were honest about what they were building and clear about which decisions should always have a human behind them.

We're asking those questions of ourselves. We'd be glad to ask them with you.