The business logic that makes your AI matter

A functioning swing hangs from two ropes. In most AI initiatives, the technology rope is well secured; that is the rope we know how to tie. It is the other rope, the one that should anchor the initiative in the business logic, that so often was never properly attached. And a swing with one rope does not swing. It tips, however strong the one remaining rope is.

An old problem meeting a radically new shift

Let us be honest: we are not looking at a new problem with AI. Connecting technology to business value has challenged organisations since long before anyone ever combined the words ‘digital transformation’, and the graveyard has been receiving well-intentioned IT projects for decades. So why write about it now?

What is new is the radical nature of the shift required, driven by a fundamental shift in the technology itself. AI has moved from chatbots to reasoners to agents that plan, act, and self-correct, with the autonomous task horizon of the best models roughly doubling every seven months. And digital and AI-native companies are disrupting the balance of established markets to a larger degree than before: they are not simply doing the same things faster, they are configured around the technology from day one – they are born digital.

The old problem now demands a more radical answer, and waiting is the one rope that no longer holds.

From the top down

So where should the answer start? Since everyone can buy the same tools, it should not start with technology. It starts with the business model – with the winners being the ones that let it pull the technology, rather than the other way around.

One picture holds that discipline together, and it has three layers.

At the top is the business model: where value is created and captured.

In the middle sits the layer that decides whether the other two ever actually connect: dynamic capabilities. Traditional companies are bombarded with inspiration on the art of the possible – consultants, vendors, case studies of Klarna, Amazon, Spotify, companies where digital and AI are the business model instead of a layer bolted onto one. That inspiration is real, but it rarely translates, because most organisations were never designed the way those companies were. Dynamic capabilities are the sensible answer to that gap: the organisation's ability to sense change early, seize the right opportunities, and reconfigure itself as conditions shift, turning a changing strategy into changed ways of working.

At the bottom are the operational and technical capabilities that deliver all of it in daily work and systems.

The pull has to run downward. The moment it reverses, and tools start searching for problems to solve, you are on the road to the graveyard.

The inflection point: a leadership ‘a-ha’ moment

What does it take, then, to move beyond productivity? The first wave of AI adoption is individual: training, tools, personal productivity. It is valuable, and something every organisation should harvest – AI will keep boosting it. But productivity is available to everyone, which is exactly why it cannot set anyone apart. It is also where most organisations get stuck: companies have invested heavily in generative AI tools that boost individual output, yet most struggle to scale beyond that, because siloed, multidivisional structures are not built to convert individual use into shared, persistent organisational capability.¹

Attaching the initiative to the business logic is what converts personal gains into organisational capability, and the six disciplines that follow are, in practice, the mechanism for realising that conversion. Plot value creation over time and the curve rises gently, then reaches an inflection point. What happens there is not technical. It is a leadership ‘a-ha’: the moment leaders stop seeing AI as a stack of use cases and start seeing the business through AI lenses. We call the capability digital imagination.

Getting past that point takes three leadership enablers:

• Imaginative competence
• Strategic prioritisation
• A redesigned operating model

The difference is simple to put into words but hard to put into effect. Before the inflection point, AI makes the organisation faster. After it, AI changes what the organisation can be.

Six disciplines of business-driven AI

Being business-driven is a practical matter – built from six connected practices:

1. Start from comparative advantage
   Everything starts with economic logic. Tools that are readily available to everyone create no advantage. Build where only you can play and utilise your existing uniqueness and make this stronger. Buy where the market already competes. And subject every initiative to hard economic tests before it is funded.
   
   
2. Lead with capability
   Skip the grand platform project. Start where the value potential is greatest, build only the foundation that the capability needs, and design it to scale, so impact is delivered at every step.
   
   
3. Build dynamic capabilities for a complex world
   Most businesses are complex, not just complicated. Complexity is full of tensions that cannot simply be solved. Resilience lives in the adaptive capabilities that sit above the technical ones. Sense opportunities and threats, seize the right ones, and transform as conditions change. Or as physicist Niels Bohr put it: "How wonderful that we have met a paradox. Now we have some hope of making progress."
   
   
4. Lead the organisation past the inflection point
   Adoption is a leadership question first, and a technology question second. Working the six disciplines is what produces the shift, visible in the questions leadership asks. Before the inflection point, the focus is on where AI can remove effort from what the organisation already does. After it, the questions change: How do we augment our current offerings? Which opportunities and threats does AI create that will reshape our clients' needs and the market itself? How does that change the value we can bring? And what would our business model look like if it were designed AI-native, for an AI-native market?
   
   
5. Anchor it in your operating model and knowledge foundation
   Know your maturity honestly and take deliberate positions on data strategy. That includes deciding how individual use becomes shared, persistent capability, rather than context that evaporates with each session. Then anchor the change in your digital operating model and the knowledge foundation beneath it.
   
   For more on this, go to: The foundation that makes your AI yours
   
   
6. Excel in leading change to gain value
   Value arrives when solutions are implemented, adopted, and harvested. That requires firmly pursuing the desired benefits through clear business ownership and dedicated change management, because even the most capable AI still needs humans in the loop to make it real.

Where to start

The business-driven AI assessment gives organisations an honest picture of where they stand across the six disciplines.

The assessment works in three stages:

1. Establish the baseline, including where leadership sits on the adoption curve
2. Choose where AI goes first, using an end-to-end view to challenge business logic and perform the economic tests
3. Build the roadmap, with owners, dependencies, and a clear benefit-realisation logic.

We propose to run it with the executive team, at group level, or for a single business unit, and because the baseline calibrates the ambition to the organisation's actual maturity, it creates value whether you are early on the curve or far along.

The roadmap covers the capability build itself – the concrete steps towards the dynamic capabilities the organisation will need – not only the AI initiatives.

It typically takes four to six weeks, and it answers the question most organisations cannot readily answer today: “Where do we start, and why there?”

... both ropes must be attached before anyone can swing.