As you approach the end of the year, management’s focus shifts to creating the plan for the next year.  Days lost in planning meetings. Budgets get fought over, deals done in corridors. Months later, the plan is complete. Winners. Losers. Certainty.

In many organisations, plans feel like control, a detailed roadmap that promises certainty in an uncertain world. The problem is that the world is uncertain, and our tendency to predict what our customers will need and lock it into our plans months, even years, ahead of time is the definition of insanity. The only certainty is that things will change. Yet we end up wasting an enormous amount of energy suppressing change, because it doesn’t fit the plan.

We plan and predict rather than test and learn. You can forecast, budget, and project all you want, but in a world that shifts constantly, plans often become relics before we start working on them.

Most organisations still have Plan and Predict built into their organisational DNA, but many now recognise that Experiment and Adapt would serve them much better. Organisations are great at deciding what they think will matter, but struggle to test what actually matters.

Plans create the illusion of control. Real control comes from learning what matters next. Instead of building organisations optimised to predict the future, we need organisations designed to test ideas quickly.

To learn what works.
And act on it.

The problem isn’t planning. It’s treating planning as a proxy for certainty.

A Different Way to Decide What Matters

What if the problem isn’t our ability to plan, but our belief that we can know what will matter in advance?

In complex, fast-moving environments, prediction breaks down. The variables are too many, the interactions too unpredictable, and the consequences too delayed. This is where a different philosophy becomes essential, an approach based on empiricism. Empiricism is about experimentation and evidence that guide our decisions, rather than designing a perfect plan up front.

Empiricism isn’t new. It’s how humans solve complex problems. We act. We observe. We adapt. Trying to predict everything in advance simply doesn’t work in complex systems because reality will always surprise us.

Yet in organisations, we often do the opposite. We try to think our way to certainty before we act.

Stepping Into the Unknown

This shift only becomes real when we take the first step. Instead of committing to our year-long plans, we might ask:

• What is the smallest thing we could do to test this idea with real customers? 
• What assumption are we making that we could validate this week? 

When we embrace the unknown, organisations cross the threshold from prediction to learning.

Stop investing in being right.
Start investing in finding out.

A small experiment replaces a large commitment. Today, with AI, that experiment can happen faster than ever. A product team can generate a working prototype in hours using AI-assisted coding, rather than waiting weeks for development. Instead of debating ideas, they can put something real in front of customers almost immediately and learn from actual behaviour.

A hypothesis replaces a requirement. A short feedback loop replaces a long delivery cycle. And something interesting happens: the conversation shifts from opinion to evidence. 

What We Learn When We Start Testing

• The first experiments rarely go as expected.
• Sometimes the idea doesn’t land with customers.
• Sometimes the impact is smaller than anticipated.
• Sometimes something completely unexpected emerges.

But this is where the real value lies, because every experiment generates evidence. AI accelerates this even further. Teams can now analyse thousands of pieces of customer feedback in minutes, surface patterns they would have missed, and validate ideas the same day. Instead of debating opinions in meeting rooms, teams begin to see what is actually happening in the market. Instead of defending the plan, they start adapting as they learn.

In an empirical approach, progress is no longer measured by how well we follow the plan. It’s measured by how quickly we learn what works and what doesn’t.

Over time, this creates a different kind of organisation. One that doesn’t need to predict the future perfectly because it has learned how to discover it, one step at a time.

What we see over and over again in organisations is that at some point, the evidence starts to contradict the plan. We’re confident we have it right this year, but every year, something happens to blow the plan out of the water.  

• Our customers ask for something different halfway through the year
• We are building a feature we thought was critical, but our latest insight tells us that nobody cares
• Our competitors are leveraging a new technology to gain a competitive edge.

We have met the real test. Increasingly, that evidence arrives faster than organisations can respond. One company I worked with used AI to prototype and test three product ideas with customers in a single week, something that would previously have taken an entire quarter. The uncomfortable truth? Two of those ideas, which stakeholders agreed upon in the annual plan, failed immediately when exposed to real users.

Do we follow the plan or follow the evidence? A hard choice for many. Because following the evidence means admitting we were wrong. And most organisations are built on being right.

But this is also the opportunity for us to make a breakthrough, the moment where organisations stop asking:

How do we deliver the plan?

And start asking:

What is the market telling us to do next?

Once organisations start learning what actually matters, a new question emerges:
How do we organise around learning, not planning?

The traditional approach batches work into projects and programmes, approved as part of a yearly planning process. Creating an all-or-nothing dynamic: either the project is delivered, or it isn’t. A more agile approach breaks work into small chunks of value or learning. What’s the smallest thing that could create impact for a customer? Or the smallest experiment that could help us learn about the market? These become investment themes, ordered in a Portfolio Backlog with the most valuable at the top.

Teams are then optimised for the flow of value, enabling fast feedback and continuous learning. Rather than funding a fixed annual plan, organisations set strategic goals and allocate capacity, while keeping the portfolio flexible. Through continuous planning, they adapt as they learn and respond to market changes.

Why This Matters Even More in the Age of AI

As organisations adopt AI, the speed of execution is no longer the constraint. Working with AI, we can generate ideas faster, write code faster, and automate decisions. But this doesn’t reduce uncertainty. It amplifies it. Because when you can act faster, you can also be wrong faster.

AI doesn’t solve the problem of knowing what matters. It makes it more urgent. In this world, competitive advantage doesn’t come from predicting the right answer upfront. It comes from learning faster than anyone else.

The organisations that win won’t be the ones with the best plans. They’ll be the ones with the fastest learning loops.

Patterns of Value Delivery to Test and Learn

Here is just a flavour of some of the enabling patterns that might allow your organisation to start delivering based on what you are learning, rather than a plan you created at the beginning of the year. 

Lean Portfolio Management

Applies lean principles to managing the portfolio of initiatives or investments, emphasising flexibility, value delivery, and learning. This approach enables organisations to adapt quickly to change, make investment decisions based on current information, and continuously align initiatives with strategic objectives.

Empirical Approach to Ordering

Utilises real-world data and feedback to decide which investments to prioritise. This approach emphasises learning from experience, adapting strategies based on observations rather than assumptions, and focusing efforts on high-impact areas and learning. We might mean using AI to continuously analyse customer behaviour, usage data, and feedback signals to inform prioritisation in near real time, shifting from periodic prioritisation to continuous, evidence-based decision-making.

Beyond Budgeting

Challenges traditional budgeting processes, advocating a more flexible, adaptive financial planning approach that responds to the dynamic business environment and encourages innovation and agility.

Value Streams

Organising teams around specific customer value to ensure that all activities are directly contributing to delivering customer value or learning. Break dependencies by having cross-functional teams focus on one investment theme at a time to accelerate learning loops and value creation.

Manage Flow

Rather than managing the plan, focus on optimising the movement of work through delivery. Effective flow management identifies and addresses bottlenecks, ensuring work progresses efficiently from start to finish and enhancing overall productivity.

Freedom to Fail

Create an environment where failure is not just accepted but seen as a crucial part of the learning process. Celebrate fast failure that results in learning, encouraging teams to make small bets, some of which might be calculated risks. When people feel safe to experiment and possibly fail, they’re more likely to propose innovative solutions and learn valuable lessons that lead to significant improvements.

What This Means in Practice

You don’t need to abandon planning. But you do need to change its role. Don’t treat planning as an annual event. Make it a continuous activity that lets you adapt as you gather evidence and learn what customers want.

It should be about creating the conditions to learn what matters next.

• Break projects into smaller chunks of value or learning. What is the smallest piece of value we could put in front of a customer? 
• Order these potential investment themes based on your strategy and other important business value drivers.
• Don’t work on everything at once; optimise for flow and deliver fast to accelerate value creation and learning. 

In a world that won’t stand still, advantage doesn’t go to those who plan best.

It goes to those who learn fastest. In the age of AI, the organisations that win won’t be the ones that move fastest. They’ll be the ones who learn what matters, before everyone else does.