The Three Horizons of AI Leadership: Adoption. Evolution. Transformation.

In Brief

After almost 30 years working as a professional futurist and leading breakthrough innovation programs for Fortune 500 companies from Microsoft to Unilever, I believe leaders need to distinguish between three fundamentally different horizons of AI value creation: AI Adoption, AI Evolution, and AI Transformation.

The latest research suggests that companies creating the greatest value from AI are moving beyond measures of activity and adoption toward workflow redesign and business evolution. The long history of disruptive technologies and innovation shows that the big wins will come from business model reinvention, and that is something corporations really struggle with.

Leaders need to work across all three horizons simultaneously, while ensuring that near-term AI activity builds capabilities and momentum toward more ambitious opportunities.

AI Activity Is Exploding. AI Value-Creation Is Much More Uneven.

Everywhere I go as an AI keynote speaker and AI leadership keynote speaker, I see organizations moving extraordinarily fast to deploy AI. Copilots are rolling out, employees are experimenting with frontier models, agents are being directed, SaaS enterprise platforms are adding AI everywhere, and boards are pushing leadership teams to demonstrate they have an AI strategy.

Much of this activity is entirely rational. Yet, as I argued in Why So Many Smart Companies Are Failing At AI Deployment, we cannot simply shove AI into existing workflows, cultures, products, and operating models and hope that radically new technology will somehow generate transformational returns from structures designed for another era.

The latest evidence is increasingly stark. PwC's 2026 research across 1,217 organizations found that just 20% of companies were capturing 74% of AI-driven economic returns. Leaders at these firms were twice as likely to redesign workflows rather than simply add AI tools, and 2.6 times as likely to report that AI was improving their ability to reinvent their business model.

BCG found something similar in July. Almost nine in ten CEOs reported some cost or revenue benefit from AI in targeted areas, yet only 14% had clearly defined P&L impact for all AI initiatives. Its high performers were roughly seven times more likely to redesign workflows and reshape the business end-to-end around AI.

The sophistication of an AI strategy should not be measured by how much AI an organization is deploying, but by the kind of value it is learning to create.

Leading Across Three Horizons of Change & Innovation

I have been working with multi-horizon thinking for almost 30 years, not only as a professional futurist but also as an innovation strategist, entrepreneur, and leader of breakthrough innovation programs for Fortune 500 companies from Novartis to Diageo to Microsoft.

Much of my work in innovation, and later coaching, training, and developing innovation leaders, has focused on transformative value creation: helping leadership teams challenge existing industry assumptions with new business and operating models, usually leveraging emerging technologies to develop products, services, and business models that deliver genuinely new sources of customer, consumer, and enterprise value.

I began using this kind of thinking before a framework called the Three Horizons, developed by Mehrdad Baghai, Stephen Coley, and David White, was published in 1999. The Three Horizons framework distinguishes between strengthening and extending the core business (Horizon 1), building emerging opportunities that can become meaningful new sources of growth (Horizon 2), and creating more transformational options that may eventually redefine the business itself (Horizon 3).

Crucially, the horizons were never intended to mean now, later, and much later. Executives are supposed to lead across all three concurrently. The key to the framework is that leaders need a way to manage today's core business while developing emerging opportunities and creating options for significant future growth.

This resonates with my own work as a futurist, ever since I trained in scenario planning with the late, great GBN. The purpose of foresight, for me, has never been to make clever predictions about the future or geek out on what is possible. It helps leaders spot weak signals of what may be emerging, challenge assumptions inherited from the past, make wiser choices in the present, and actively shape a preferable future for themselves, their customers/consumers, and the world itself.

For the AI age, I think about the 3 Horizons as AI Adoption, AI Evolution, and AI Transformation.

Horizon 1: AI Adoption

AI Adoption means integrating AI into activities, tools, products, services, and roles we already have so they become faster, cheaper, smarter, or more effective. It encompasses much of today's deployment: copilots, content generation, customer service chatbots, coding support, automated analysis, search, individual productivity, and AI features added to existing products.

This horizon matters enormously. It builds AI fluency, creates practical learning, and can generate significant productivity, quality, and efficiency gains.

It also connects directly with my framework of The LEADERSHIP-AI Synthesis. Rather than either rejecting AI or outsourcing our agency and judgment to it, we learn to collaborate with machine intelligence that stimulates, analyzes, challenges, generates, and extends human thinking.

But AI Adoption still largely starts with the business we already have. It asks: How can AI help us do what we already do better?

Deloitte's recent research shows the limitations of staying in this horizon and mistaking successful AI deployment for successful AI transformation. Forty-eight percent of respondents said their organizations introduced AI without redesigning workflows or roles around it, while only 12% reported redesigning at scale with a new operating model. Copilots deployed, employee access/logins, and token usage are poor proxies for transformation.

Horizon 2: AI Evolution

AI Evolution occupies the critical territory between embedding a new technology into Business-As-Usual and genuinely transforming the business.

With evolution, we are still drawing heavily on assets we already possess: our technologies, expertise, customer relationships, brands, products, services, data, distribution, and organizational capabilities. But we no longer accept how those assets are currently assembled as inevitable.

We rewire workflows. We rethink how humans and agents collaborate. We collapse unnecessary handoffs. We change decisions and roles. We evolve existing products and services around intelligence that was previously unavailable. We reconfigure how value travels across organizational silos.

Imagine an existing workflow with 18 stages crossing six functions. Horizon 1 might use AI to speed up stages three, nine, and fourteen. Horizon 2 asks whether, if predictive AI, generative AI, agents, automation, and human judgment had existed when the workflow was originally designed, anybody would have created those 18 stages in the first place.

Harvard Business Review's latest piece on agentic AI illustrates how significant this may become. Researchers studying organizations including Walmart, Amazon, Ericsson, Ramp, and Medtronic argue that one of the next frontiers is orchestrating work across functional silos.

With such AI evolution, AI can increasingly perform analyses, route information, and surface trade-offs, while humans contribute tacit knowledge and context, set guardrails, and make consequential judgments.

The same principle applies externally. An existing service may become radically personalized. A software tool may evolve into an agent that completes an outcome. An insurer may move from responding to events toward anticipating and preventing them. Existing propositions can become significantly more valuable without yet becoming an entirely new business.

This is why I prefer evolution to scaling. Scaling tells us how widely something is being deployed. Evolution tells us that the system itself is changing.

Horizon 3: AI Transformation

Transformation is a big word, and I reserve it for a big phenomenon. AI Transformation begins when leaders question not simply how their organization works, but some of its deepest assumptions about what business it is in, what value it creates, for whom, and how.

What becomes possible when intelligence that used to be scarce becomes abundant? What entirely new product, service, customer experience, category, or business model could emerge? What existing sources of competitive advantage become irrelevant? What unmet customer or human need suddenly becomes economically viable to solve?

This is where my idea of Adaptive Intelligence becomes as important as Artificial Intelligence. Artificial Intelligence can help us generate extraordinarily sophisticated answers. Adaptive Intelligence helps us recognize when reality has changed so profoundly that we are asking the wrong questions.

AI is dramatically expanding the strategic possibility space. AI can overcome some of the traditional limits that time, complexity, and human cognitive bandwidth place on strategic innovation by helping leaders generate, interrogate, evaluate, and synthesize many more strategic possibilities.

This makes human leadership more important, not less. As I explore in my recent work on Contextual Judgment in the AI age, AI can expand what we can see and suggest what might happen. Leaders remain responsible for determining what ought to happen in this specific human, ethical, commercial, cultural, and strategic context.

Defining the Problem Right Changes Everything

One of the most consequential things I have learned through decades of breakthrough innovation work is that generating ideas is rarely the hardest part. The real work is defining the right problem, opportunity, or unmet need to innovate around.

This takes time, attention, reflection, reframing, mindset expansion, customer insight, challenge, imagination, and significant energy. The obvious problem is frequently not the most valuable one because the initial brief is usually framed by assumptions inherited from Business As Usual.

This is why I argue that leaders need to think outside the box before they build a better box.

Get the problem wrong, and even brilliant people with brilliant AI can generate hundreds of ingenious answers to an inappropriate question, one stemming from the comforts of the past, not the challenges of the future.

Get the problem right, and something remarkable starts to happen.

In my experience, a powerful Horizon 3 challenge may reveal one or two genuinely transformational opportunities. Those can generate perhaps three or four compelling Horizon 2 opportunities to evolve existing products, services, workflows, and operating models, which in turn can surface nine, ten, or more useful Horizon 1 opportunities.

The ratios are obviously not mathematical laws. The pattern is what matters. Horizon 3 gives the other horizons direction toward becoming future-proof, even future-positive (becoming more valuable as the future rushes towards us).

Instead of asking thousands of employees to find random places to put AI, leaders establish a meaningful future toward which today's innovation can begin moving.

Horizon 3 provides the direction. Horizon 2 creates the bridge. Horizon 1 creates momentum and readiness.

Use AI Adoption to Start Transitioning Toward AI Transformation

This changes everything about how we choose near-term AI adoption opportunities.

The strongest adoption initiatives can generate immediate value while building the capabilities, data, behaviors, expectations, worldviews, confidence, and organizational muscle required for AI Evolution and, eventually, AI Transformation.

Imagine that your AI Transformation vision involves moving from a standardized service to an intelligent, continuously personalized customer experience that replaces your existing service with one that solves a problem that is becoming more painful for more people, as the future unfolds.

AI Evolution might require redesigning the customer journey, changing how human expertise and agents interact, or evolving an existing product. That then reveals AI Adoption experiments and initiatives you can execute now: gathering new kinds of data, offering modest personalization, introducing decision-support tools, experimenting with agents, changing customer communications, or helping employees learn to collaborate effectively with Alternative Intelligence.

Customers start to experience elements of the future proposition. Employees develop new behaviors and ways of working. Processes evolve. Governance matures. Leadership learns. The organization progressively becomes capable of holding the transformation before the business model innovation ever launches.

This is much less risky than developing a brilliant transformation in isolation and then discovering that the existing organization, its customers, or its people are simply not ready for it.

This Is Fundamentally an AI Leadership Challenge

Technology teams can deploy AI. They cannot decide, on their own, what the organization should become because AI exists.

That requires executives who can hold all three horizons simultaneously: exploiting what works today without becoming imprisoned by it, evolving the organization as new possibilities emerge, and protecting enough time, imagination, courage, and investment to explore genuine transformation.

This is increasingly what I seek to unlock in my work as an AI leadership keynote speaker.

A highly customized AI leadership keynote can help bring tens, hundreds, or even thousands of leaders into coherence on what is possible and what is needed to respond to this unprecedented disruption with the appropriate level of ambition and action.

It can raise the altitude of the conversation from "Where else can we put AI?" toward "How should we evolve our operating model?" and ultimately "What could we become and why?"

It can also get executives in the right state of mind and body to ask uncomfortable questions and define the real problems AI could solve.

From AI Keynote to Transformative Action

Where appropriate, I can then build on the keynote with a carefully designed interactive workshop, using the shared energy and language created in the keynote to apply this thinking to real business, customer, leadership, and AI transformation challenges.

I have developed this format over many years as both a keynote speaker and facilitator. You can explore how I combine a keynote with a deeper interactive working session in Beyond the Keynote and Into the Workshop and see the different keynote and workshop formats I have pioneered.

The purpose is not to produce another wall covered with 97 AI use cases. The intent is to help leaders frame the problem right, identify a small number of high-value opportunities, and work backward to surface the AI Adoption and AI Evolution activities that could start creating AI Transformation readiness now.

For senior teams that need more time and protected space to do this properly, my leadership consultancy, Switch On Leadership, designs and facilitates executive retreats and leadership offsites to help leaders step away from Business As Usual, relax enough to stop stress-driven blind spots and habits from undermining future-forward thinking, challenge assumptions in a brave yet safe space, explore possible futures with enough time to reflect on what it means, prioritize high-value opportunities and investments, and align around AI-enabled business transformation (and what it might take from them).

In a companion article to this piece, How Executive Retreats and Leadership Offsites Can Accelerate Effective AI Adoption, AI Evolution, and AI Transformation, I focus on the architecture of an offsite or retreat designed to ensure leaders engage in the challenge with their full capacities and collective intelligence.

For organizations that want to go deeper, Switch On Leadership designs and delivers customized, experiential AI leadership development programs that build the AI fluency, Adaptive Intelligence, Contextual Judgment, creativity, relational capability, and transformational leadership required to turn AI investment into sustained business impact.

AI Transformation for Optimal Value-creation

AI is too consequential to be treated merely as a technology rollout. It can make today's activities considerably more effective. It can help us evolve existing workflows, products, services, and operating models. And, most consequential of all, it can enable leaders to imagine forms of value creation that were simply not possible before.

Great AI leadership means doing all three, while understanding which you are working on and why.

Embed AI into the business you have. Evolve the business around what AI enables. Transform the business for the world AI is creating.

If you'd like to work with me as your AI keynote speaker, get in touch with my team to check dates.

FAQs On AI Transformation

What are the Three Horizons of AI Leadership?

The Three Horizons of AI Leadership are AI Adoption, AI Evolution, and AI Transformation. AI Adoption improves activities within today's business. AI Evolution rewires existing workflows, products, services, and operating models around what AI enables. AI Transformation explores fundamentally new ways for an organization to create customer, enterprise, and strategic value.

What is the difference between AI adoption and AI transformation?

AI adoption integrates AI into activities, products, and workflows an organization already has. AI transformation goes much further by questioning underlying assumptions about what the organization does, how it creates value, whom it serves, and what new products, services, experiences, or business models AI could make possible.

What is AI Evolution?

AI Evolution is the critical middle horizon between adoption and transformation. Organizations continue using many of their existing assets, capabilities, products, services, and customer relationships, but substantially rewire workflows, decisions, human-agent collaboration, operating models, and propositions around what predictive, generative, and agentic AI now enable.

How should companies measure AI value?

Companies should measure different forms of value across different horizons. AI Adoption can be measured through productivity, effectiveness, quality, and cost. AI Evolution should increasingly produce measurable business outcomes such as revenue, margin, customer value, cycle time, and decision quality. AI Transformation should create strategic value through new propositions, markets, revenue streams, and business models.

Why is AI transformation a leadership challenge?

Technology teams can deploy AI, but they cannot determine alone what an organization should become because AI exists. Leaders must decide what value to create, challenge inherited assumptions, work across multiple horizons, exercise contextual judgment, mobilize people through change, and determine which capabilities and decisions should remain profoundly human.

What can an AI leadership keynote help an organization achieve?

A customized AI leadership keynote can rapidly create shared understanding across an executive or leadership population, distinguish AI activity from AI value, introduce the Three Horizons of AI Leadership, and raise the conversation from “Where else can we deploy AI?” toward “How should our business evolve?” and ultimately “What could we become?”