AI Is Not a Strategy
AI is everywhere.
It is embedded in products, workflows, search tools, analytics platforms, customer experiences, software development, decision support, and increasingly the everyday work of employees.
Organizations are investing in it.
Leaders are talking about it.
Roadmaps are being built around it.
And increasingly, organizations describe AI itself as part of their strategy.
But there is an important distinction worth making.
AI can enable a strategy.
It can dramatically change how an organization executes a strategy.
It can create capabilities that were previously unavailable.
But adopting AI is not, by itself, a strategy.
Strategy begins somewhere else.
It begins with what the organization is trying to accomplish.
Start With the Outcome, Not the Tool
Imagine an organization with a large product catalogue. Customers frequently purchase the same products, but the timing varies. The organization has years of transaction data and can identify patterns in purchasing behaviour.
One possibility is to introduce an AI-enabled capability that recognizes those patterns and proactively suggests when a customer may need to reorder.
That may be an excellent application of AI.
But:
Use AI to recommend repeat purchases.
is not the strategy.
The strategic intent might be:
Anticipate recurring customer needs and reduce friction in the purchasing experience.
Now the role of AI becomes clearer.
It is a capability that may enable the organization to accomplish something it has already decided matters.
The distinction may sound semantic.
It isn't.
Because the order in which we ask these questions changes the way organizations invest.
Tool-First Thinking Reverses the Sequence
When AI becomes the starting point, the organizational question becomes:
Where can we use AI?
That question will almost always produce answers.
AI is extraordinarily versatile.
It can summarize information.
Generate content.
Identify patterns.
Automate activities.
Support decisions.
Search large bodies of information.
Interact with customers.
Accelerate analysis.
There are likely hundreds of places in even a modest organization where AI could be used.
But could we use AI here? and would using AI here improve something that matters? are very different questions.
Tool-first thinking can lead organizations to select the intervention before they have diagnosed the problem.
A more useful sequence begins with intent.
What are we trying to accomplish?
Then diagnosis.
What is preventing us from accomplishing it today?
Then intervention.
What needs to change?
Only then does the technology question become useful:
Can AI help us do this better?
Sometimes the Problem Is Not a Technology Problem
Suppose a process takes too long. AI may be able to automate parts of it. But why is the process slow?
Perhaps employees spend hours manually synthesizing information.
AI could be extremely useful.
Or perhaps work waits three days for an approval nobody actually needs.
Perhaps information has to be entered into several disconnected systems.
Perhaps responsibility for the decision is unclear.
Perhaps people routinely redo work because the required output was never clearly defined.
Perhaps the process itself no longer serves a useful purpose.
Automating any of those activities might make the process faster.
It could also make a poorly designed process operate faster.
This is why technology cannot replace diagnosis.
Before asking how AI can improve a workflow, it is worth understanding what within that workflow actually needs to improve.
AI Can Be an Enabler of Strategy
None of this diminishes the strategic importance of AI.
Quite the opposite.
AI may enable organizations to do things they previously could not do, or to do existing things at dramatically different levels of speed, scale, cost, or personalization.
A customer strategy may use AI to better anticipate needs.
An operational strategy may use AI to reduce repetitive work.
A growth strategy may use AI to identify patterns or opportunities that were previously difficult to detect.
A capability strategy may use AI to augment how employees access and apply organizational knowledge.
A product strategy may incorporate AI into features that create new customer value.
In each case, AI can materially affect the organization's ability to execute.
But the strategic question remains:
What are we trying to achieve?
AI answers a different question:
What might help us achieve it?
Even an AI Strategy Needs an Organizational Purpose
There is a legitimate reason organizations use the term AI strategy.
Organizations need to make deliberate choices about how they will build, buy, govern, secure, integrate, and deploy AI capabilities.
They need to consider data.
Risk.
Architecture.
Capability.
Governance.
Investment.
Workforce implications.
Those choices matter, and collectively they may reasonably be described as an AI strategy.
But even that strategy needs a parent.
Why is the organization developing these capabilities?
What organizational outcomes are they intended to enable?
Where does AI create meaningful advantage or improvement?
Where does it not?
Without those answers, an AI strategy can become a technology roadmap disconnected from organizational intent.
The existence of an AI strategy does not answer the more fundamental strategic question:
What are we trying to accomplish?
The Broader the Tool, the More Important the Diagnosis
AI presents an unusual organizational challenge precisely because it is so broadly applicable.
When a technology can potentially be used almost everywhere, finding somewhere to use it tells us very little. The broader the possible applications of a technology, the more important it becomes to define the problem before selecting the application. Otherwise, organizations can spend enormous amounts of energy implementing AI while remaining unclear about what improvement those implementations are intended to produce. And once that happens, another problem appears.
How do we know whether any of it worked?
If the objective was simply use AI, adoption itself can begin to look like success.
More users.
More licenses.
More automated tasks.
More AI-enabled features.
More pilots.
More output.
Those measures may tell us something about adoption.
They tell us much less about organizational value.
Strategy Gives AI Something to Accomplish
The point is not to slow AI adoption until every possible consequence has been analyzed.
Organizations sometimes need to experiment.
They need to learn.
They may need to move quickly as technology and competitive conditions change.
But experimentation can still have intent.
What are we trying to learn?
What constraint are we testing?
What improvement are we looking for?
What happens elsewhere in the system if this activity becomes faster?
What would tell us that this application is worth expanding?
Those questions do not inhibit experimentation.
They give it purpose.
AI is an extraordinarily powerful tool. That power may fundamentally change what organizations can accomplish and how they accomplish it.
But power does not determine purpose.
Before asking:
How should we use AI?
there is a more important question:
What are we trying to achieve?
Only then can we determine whether AI is the right tool to help us get there.
Related Reading
AI as a Systems Stress Test: Explore how AI adoption can expose weaknesses in workflows, governance, decision rights, information flows, and organizational capability.
What Kind of Problem Are We Actually Looking At?: Explore why selecting an intervention begins with understanding the problem rather than responding only to what is immediately visible.
Seeing Systems Instead of Symptoms: Explore why organizational outcomes often emerge from interactions across a system rather than a single identifiable cause.
Related Frameworks
Organizational Intentionality: Explores how deliberate choices and accumulated decisions shape the organization over time.
Decision Altitude Model: Explores how broadly a situation needs to be understood before determining what intervention is appropriate.
Comments