Not all AI transformations are the same

Not all AI “transformations” look the same.

Most leadership teams I meet are already determined that AI transformation is a priority. Whether it’s pressure from their board, or their CEO/CIO, they all know they need to drive adoption of AI in their business.

The question they can't answer, at least not as a team, is what exactly they're trying to transform, and how they expect that change to actually happen.

That gap matters more than people realise.

I've seen leadership teams spend months debating AI strategy without ever realising they were describing completely different destinations. One executive is thinking about automating a handful of operational tasks. Another is imagining rebuilding entire workflows around AI. A third is quietly hoping to use AI to enter new markets entirely.

This isn't a communication problem. It's a framing problem. And without the right frame, even well-intentioned AI investments fragment before they compound.

The two questions every leadership team needs to align on

When I work with leadership teams on AI strategy, I focus on two dimensions: The first is what you're transforming. The second is how you plan to make the change happen.

Let’s go through each dimension on its own before putting them together for the bigger picture.

The "what": four levels of AI ambition

Not all AI initiatives are created equal. There's a meaningful spectrum, and where you're aiming has significant implications for how much investment, change management, and leadership commitment is actually required.

The question isn't which level is best. It's which level you're actually aligned on, and whether your investment, timeline, team enablement and risk appetite match it.

Point solution automation is where most companies start: AI handles a specific, well-defined task: writing first drafts, summarising meeting notes, processing documents. The gains are real but bounded.

Human augmentation is a step up in ambition. Here, AI becomes a co-pilot across the workforce, tackling lower-value work, elevating output quality, and freeing people to focus on what actually requires human judgment.

Core process reinvention means rebuilding workflows around AI — not layering it on top of existing processes, but redesigning how work actually gets done. Roles change. Decision rights shift.

Business model transformation is the rarest and most demanding level. This is where AI doesn't just improve how you do things — it changes what you compete on.

The "how": four change approaches

Ambition is one thing. Change approach is another. This is where I see the most dangerous misalignment.

Top-down mandate is exactly what it sounds like: the CEO or COO declares direction, picks the domain, and assigns the organisation's best people to it.

Capability building is a longer game. The focus is on foundations first: AI literacy, data infrastructure, governance, internal talent. Transformation emerges from the capability you build, rather than being declared from the top.

Lighthouse and scale is where I see the best return for most mid-market organisations. Pick one domain. Go deep. Prove visible impact. Use that proof point to unlock resources and replicate the model. The discipline here is in the selection: a lighthouse that's too narrow doesn't prove enough; one that's too broad doesn't prove anything.

Grassroots adoption is where most Australian companies actually are right now. AI spreads organically from individuals and teams doing their own thing — buying their own tools, building their own prompts, finding their own shortcuts. The individual gains are real. The problem is that they rarely compound into enterprise value. BCG's research is clear: 74% of companies struggle to achieve and scale value from AI — and the leaders who do consistently focus 70% of their effort on people and processes, not technology. Grassroots adoption is a valid starting point. It's not a destination.

Putting it together: the AI Transformation Matrix

The matrix maps these two dimensions together. Each cell represents a combination of ambition and approach — some are natural fits, some are viable with the right conditions, and a few are genuinely difficult or rare.

Most mid-sized Australian organisations land in the bottom-left: grassroots adoption, point solution focus. Real wins, but no compounding impact.

How to use this matrix

There's no universally correct position on this grid. The right combination depends on where your organisation is today, what your risk profile looks like, and how much genuine leadership capacity you have to drive change.

But there are some things you can't do — regardless of ambition or resources.

You can't skip steps.

Organisations that jump straight to business model transformation without the capability foundation tend to create chaos rather than value. The progression is real: you need to build the muscle before you can use it at scale.

This isn't a one-time destination.

Reaching augmentation is a milestone, not an endpoint. The organisations compounding their AI advantage are treating each stage as a foundation for the next.

You can't get to business reinvention through a purely bottoms-up approach.

McKinsey's research is clear: nearly two-thirds of organisations are still stuck in pilot or experiment mode, and only 6% achieve meaningful financial impact from AI.⁴

The gap almost always comes back to the same two factors — leadership alignment and workflow redesign. High-performing organisations are nearly three times more likely to have senior leaders demonstrating clear ownership and long-term commitment to AI, and nearly three times more likely to have fundamentally redesigned workflows rather than layering AI on top of old processes. Grassroots adoption produces individual wins. Structural reinvention requires structural commitment.

Strong, fluent leadership isn't optional.

Not leadership that's enthusiastic about AI; leadership that genuinely understands what AI can and can't do, and is willing to make the hard calls about where to focus and where not to. That's rarer than it sounds.

Start with the conversation, not the roadmap

If you take one thing from this framework, it's that the matrix is most useful before you build a strategy, not after.

Use it to surface where your leadership team actually agrees or disagrees:

• Where is everyone pointing at "augmentation" but secretly imagining different things?
• Who thinks you're already doing lighthouse-and-scale when you're actually doing grassroots?
• What investments are currently in flight and what's the gap between those investments and the ambition you're claiming?

The goal isn't to make it to the upper-right corner. It's to get the leadership team genuinely aligned on what you're building toward and how you plan to get there.