Epilogue
Good strategy needs a process. Great strategists know when to break it.
I have made a strong case throughout this collection for sequence, connected tasks and explicit decisions. There is a danger in making that case too successfully. Readers could conclude that good strategy is what happens when people follow a disciplined process without deviation.
It is not.
Process protects the work from predictable failure. It makes sure the brief is interrogated, the diagnosis exists, evidence has status and choices survive downstream. Great strategists also know when the process has stopped helping.
Process is a map, not a railway line
A useful process tells you which questions matter and how decisions depend on one another. It does not require every assignment to move through the same boxes at the same speed.
Sometimes the competitor analysis reveals the answer before the audience work is complete. Sometimes a creative idea reframes the problem. An analogy from another category can expose a possibility the approved framework would never generate. A strategist may jump ahead, return upstream or ignore a tool that adds no value in this case.
The discipline lies in understanding the consequences. If the team starts with an idea, it should still test whether the idea addresses the problem, fits the audience situation and belongs to the brand. Breaking sequence does not remove the dependencies.
Serendipity needs something to collide with
The romantic account of strategy gives all the credit to the spark. The procedural account gives it to the method. Both are incomplete.
Lateral thinking often appears because a mind full of evidence, experience and unresolved tension encounters something unexpected. The process creates material for collision. It cannot schedule the insight.
This is one reason complete automation is unconvincing. A system can generate analogies and unusual combinations, sometimes brilliantly. A person still needs to recognise which surprise opens the problem and which is merely novel.
Fresh eyes can be more valuable than memory
Persistent context improves continuity and can ossify assumptions. A team that knows the client well may stop seeing alternatives because every new question arrives preloaded with the history of what was rejected and what stakeholders usually approve.
There are moments when the right move is to bracket part of that memory. Ask how the problem looks without the inherited positioning. Give somebody new enough context to understand the stakes and little enough to question the folklore. Revisit a discarded route because the market or organisation has changed.
Fresh eyes are not ignorance. They are a deliberate challenge to context that has become too comfortable.
Standardised process can compress diversity
If many agencies use similar AI workflows, strategic work may become more competent and more homogeneous. The same frameworks, evaluation criteria and model behaviours can pull teams toward the same sensible answers.
This does not make process undesirable. It makes divergence a design requirement. Bring in contradictory evidence. Use analogies from distant categories. Ask what the process is systematically unable to notice. Let people pursue a route before they can fully justify it, then subject it to the normal tests.
Raising the floor should not require lowering the ceiling.
Know why you are breaking it
There is a difference between productive deviation and avoiding the difficult part. Skipping audience work because a powerful idea has emerged may be justified. Skipping it because the deadline is close is ordinary compression dressed as intuition.
The strategist should be able to say which step is being bypassed, why, what risk this creates and how the result will be tested. That keeps flexibility accountable without turning it back into bureaucracy.
Good process catches predictable mistakes and preserves decisions. Great strategists use it as a support, notice when it narrows the answer and break it consciously. Then they return to the underlying questions to see whether the leap survives.