5 shifts shaping the adaptive business

Predicting the future is a dangerous business.

Every year brings another collection of trends that will apparently transform everything. Some do. Most don’t. And many of the changes that eventually matter most spend years developing quietly before suddenly appearing obvious in hindsight.

So I don’t find predictions particularly useful.

What I do find useful is looking for shifts that are already happening and asking what they might mean if they continue.

Right now, several of those shifts are beginning to collide. AI is moving from something people use to something capable of acting. Customers are beginning to delegate decisions to their own technology. Geopolitics is changing assumptions about supply chains, technology and markets. Trust is becoming something that needs to be designed into products and systems. And the AI conversation itself is beginning to move beyond productivity towards a much larger question of value creation.

Individually, each is interesting.

Together, I think they point towards something more fundamental.

Businesses have traditionally been designed around a reasonably stable set of relationships between customers, people, processes, technology, suppliers and economics. Those relationships are becoming less stable.

That makes adaptability more than an organisational virtue.

It is becoming part of how we need to design the business itself.

I call that The Adaptive Business.

Not a business that constantly reorganises itself or chases every new technology. And not simply another name for agile.

An Adaptive Business is one designed to continuously sense change, redesign how it creates value and turn new possibilities into measurable progress.

Five shifts are making that capability increasingly important.

01 — AI is moving from tool to actor

For the last few years, most of us have experienced AI as something we use.

We ask it a question. It writes something, analyses information, produces code or helps us complete a task faster.

That relationship is beginning to change.

AI agents can increasingly perform sequences of activities rather than individual tasks. They can interpret an objective, determine what needs to happen next, interact with systems and execute parts of a workflow with varying levels of human involvement.

That sounds like a technological distinction, but I think its implications are organisational.

A tool fits into an existing process.

An actor can change the process.

If an AI system can qualify a sales opportunity, gather customer information, prepare a proposal, coordinate internal approvals and initiate the next action, we are no longer simply making an employee more productive. We are changing how work moves through the organisation.

McKinsey’s latest research describes AI — particularly agentic AI — as requiring changes to operating models because it alters how decisions are made, how work crosses functions, how capabilities develop and where value is created.

That raises much more interesting questions than Which AI tools should we buy?

What should people continue to own? What can an agent own? Where should decisions move between them? What knowledge does the agent need? What authority should it have? When should it stop and ask a human? And who remains accountable?

Those are not simply technology questions.

They are Business Design questions.

The next generation of AI transformation may therefore be less about giving everybody an AI assistant and more about redesigning work around a new combination of people and intelligent systems.

02 — Your next customer might have an agent

The second shift is happening on the other side of the business.

Customers are beginning to use AI too.

For the last twenty years, companies have invested enormous amounts of money learning how to appear in search results, build digital customer journeys, optimise conversion and create brands people remember when they are ready to buy.

AI agents potentially insert a new participant into that relationship.

Imagine asking an agent to find the best insurance policy for your family, choose a hotel for a weekend in Rome, renew your electricity contract, identify the right software for your company or buy replacement running shoes.

The agent can potentially compare alternatives, understand preferences, interpret terms, evaluate reviews and eventually complete the transaction.

The customer may never visit most of the companies being considered.

Accenture’s 2026 research among 25,590 consumers across 16 countries found that 74% said they would trust a personal AI agent more than their best friend to make a purchase on their behalf. Accenture argues that brands increasingly face two audiences: the person and the agent acting for that person.

Whether adoption ultimately happens at exactly that pace is less important to me than the direction of travel.

For decades, brands have primarily designed propositions for people.

Increasingly, those propositions may need to be understood by people and machines.

An agent doesn’t necessarily care that your advertising is beautiful. It may care whether your price is transparent, whether the proposition matches its user’s requirements, whether delivery is reliable and whether your claims can be verified.

That doesn’t make brand irrelevant. People still have emotions, identities, preferences and relationships with brands.

But it could change where competitive advantage is created.

Companies may increasingly need to ask not only:

Why should a customer choose us?

but also:

Why should their agent choose us?

03 — The AI conversation is moving from efficiency to value creation

The first economic argument for almost every major technology is efficiency.

AI has been no different.

How many hours can we save? How much can we automate? How much faster can we complete the process? How much can we reduce cost?

Those are legitimate questions. There are enormous efficiency opportunities across most organisations.

But they are only part of the potential.

The more interesting question is:

What can we make better?

And beyond that:

What can we now create that wasn’t possible before?

I think about AI’s impact on a value chain in three simple ways: better, more efficient and new.

Better means improving the value already being created: better customer experiences, higher quality, stronger decisions, greater precision, more personalisation or giving people better capabilities with which to do their work.

More efficient means improving how that value is produced: reducing unnecessary work, increasing capacity, simplifying processes, shortening cycle times or lowering cost.

New means using AI to create something the business couldn’t realistically create before: a new capability, experience, proposition, service or business model.

The categories overlap. That is exactly the point.

If we analyse customer service purely through the lens of cost, we may discover opportunities to automate interactions and reduce handling time.

Look at the same part of the value chain through value creation and we may discover that AI can give employees dramatically better access to knowledge, improve service quality, anticipate customer needs, personalise interactions, prevent churn and identify new commercial opportunities.

McKinsey’s 2026 research similarly finds that capturing AI value is tied to redesigning the operating model rather than simply adding technology to the existing organisation.

Technology adoption eventually becomes normal.

The advantage comes from what we redesign around it.

The companies that create the greatest value from AI may therefore not be those deploying the most AI.

They may be those that become best at asking:

Where in our value chain could this make us meaningfully better?

04 — Resilience is becoming adaptability

For several decades, much of business design has been built around optimisation.

Global supply chains became extraordinarily efficient. Activities moved to places where they could be performed most economically. Inventory was reduced. Just-in-time systems removed slack. Companies became increasingly interconnected across countries, suppliers and technology platforms.

Efficiency created enormous value.

It also created dependencies.

Pandemics, wars, trade disputes, supply-chain disruption, energy shocks and changing regulation have repeatedly exposed them.

The World Economic Forum’s 2026 Global Risks Report ranks geoeconomic confrontation as the leading immediate global risk and describes trade, finance, technology, sanctions and supply chains increasingly being used as instruments of geopolitical competition.

That isn’t simply a risk-management problem.

It changes how businesses should be designed.

Which suppliers do we depend on? Where does our technology come from? Where does our data sit? How quickly could we move production? What happens if access to a market changes? Which capabilities should we own and which should we access through partners?

The most efficient system under one set of conditions can become remarkably fragile when those conditions change.

So I think the next step beyond resilience is adaptability.

Resilience asks:

Can the business withstand change?

Adaptability asks:

Can the business change?

That difference matters.

Products can become more modular. Supply chains can include alternatives. Technology architectures can reduce dependencies. Partnerships can expand and contract. Capabilities can move across organisational boundaries. Resources can be redirected as circumstances change.

The challenge isn’t designing the perfect business for today’s conditions.

It is designing a business capable of becoming something different when those conditions no longer apply.

05 — Trust is becoming part of the product

Trust has always mattered in business.

But traditionally it could often be treated as something surrounding the product: brand reputation, customer service, security, compliance and corporate behaviour.

Technology is increasingly pulling trust into the product itself.

Consider an AI agent acting on your behalf.

To use it meaningfully, you may give it access to email, documents, company systems, customer information or financial data. You need to trust what it does, understand the authority it has and know who remains accountable for its actions.

The same issue appears inside businesses.

As agents move from observing and recommending to actually acting, governance can no longer be binary. Gartner argues that different levels of agent autonomy create different trust boundaries and therefore require different governance.

The dynamic also appears in customer relationships.

If an AI agent is comparing products for a customer, claims need to be verifiable. Pricing needs to be understandable. Data needs to be structured. Terms need to mean what they say.

Trust becomes something the business has to design into the experience.

And perhaps that changes the way we think about governance.

There is a temptation to see governance as the thing that slows innovation down. Sometimes it does.

But the opposite can also be true.

If customers, employees and partners trust a system, they may be willing to give it greater authority. Greater authority may allow it to create more value.

Trust, then, isn’t simply a constraint on innovation.

It can become an enabler of it.

The Adaptive Business

These five shifts appear different.

AI actors change work. Customer agents change commerce. AI changes value creation. Geopolitics changes value chains. Trust changes the design of products and systems.

But they share something.

They make the relationships inside and around a business less fixed.

And that is why I think we need to move beyond the idea that a business is something we design, optimise and then periodically transform when the world changes enough to force us to.

Perhaps the ability to adapt needs to be designed into the business from the beginning.

A business designed to continuously sense change, redesign how it creates value and turn new possibilities into measurable progress.

This doesn’t mean changing everything all the time.

An adaptive business still needs a strong core. It needs clarity, consistency, operating discipline and capabilities that compound over time.

But around that core it needs the ability to understand what is changing, decide what matters, redesign parts of its value creation system and learn from what happens next.

I see four capabilities at the heart of that.

Sense: understand what is changing across customers, markets, technology, competition, society and the wider business environment. Not every signal deserves a response. The capability is partly about separating meaningful change from noise.

Design: translate relevant change into implications for the business. What could this mean for our proposition? Our customers? Our value chain? Our processes? Our people? Our technology? Our economics? And what might we now do differently?

Make: turn possibilities into something tangible. A proposition. A prototype. A new workflow. An AI agent. A customer experience. A service. A new operating model. Making allows the organisation to move from debating possibilities to learning from something real.

Adapt: use evidence to change the business. Scale what works. Improve what almost works. Stop what doesn’t. Reallocate resources. Update the operating model and feed what has been learned back into the next cycle.

Then sense again.

The point isn’t the arrows.

The point is that adaptation becomes continuous rather than episodic.

That, to me, may be one of the defining capabilities of the next generation of businesses.

We don’t know exactly what business will look like five years from now.

That isn’t a particularly useful prediction to make.

But we can already see enough to know that the businesses most likely to thrive won’t simply be those designed best for the world as it exists today.

They will be the ones that become very good at designing what comes next.