AI can produce a strategic plan before a leadership team has finished its first coffee.
That is not the same as producing a strategy.
A polished document can hide weak assumptions, unresolved tradeoffs, and an incomplete understanding of the buyer. In fact, speed can make those problems harder to see. Once an answer arrives in confident language, teams tend to debate the wording instead of questioning the evidence underneath it.
The real opportunity is not to ask AI to make strategy for us. It is to use AI to make our strategic thinking more rigorous: surface assumptions, organize evidence, widen the option set, challenge comfortable conclusions, and reveal where we still need human judgment.
That distinction matters because strategy is not a summary of what a company knows. Strategy is a set of choices about where to compete, which buyers to serve, what value to create, what not to pursue, and which risks leadership is willing to own.
Strategy is a system of choices—not an AI-generated document
Generative AI is exceptionally good at producing the artifacts that surround strategy. It can draft a market analysis, summarize interviews, compare frameworks, propose scenarios, build a presentation, and turn a workshop transcript into an action plan.
Those outputs can save enormous amounts of time. But an attractive output can still be strategically empty.
A real strategy forces a company to make choices that cannot all be optimized at once. Serving one buyer may mean disappointing another. Building a differentiated capability may require declining near-term revenue. Entering a market may expose the company to operational risk. A model can describe those tradeoffs, but it cannot accept responsibility for them.
This is why Harvard Business Review’s warning about AI weakening original leadership judgment is so relevant. When leaders begin with the model’s answer, they risk narrowing their own perception and interpretation before they have formed an independent view.
The goal is not less AI. The goal is a better division of labor between AI and leadership.
| Strategic work | AI can contribute | Leadership must own |
|---|---|---|
| Understanding the situation | Organize research, identify patterns, summarize evidence, and surface contradictions | Determine which evidence is credible, material, and relevant to the decision |
| Creating options | Generate scenarios, alternatives, implications, and questions the team may have missed | Decide which possibilities fit the company’s ambition, capabilities, and appetite for risk |
| Evaluating choices | Compare assumptions, model consequences, and pressure-test a decision against buyer perspectives | Make the tradeoff and explain why this choice is better than the alternatives |
| Executing strategy | Translate choices into plans, briefs, measures, and coordinated workflows | Assign accountability, resolve conflict, change resources, and learn from results |
Four jobs AI can do well in strategic planning
1. Synthesize
AI can compress a large body of customer research, market evidence, operating data, and internal knowledge into a usable starting point. The value is not merely summarization. A good synthesis preserves disagreement, source, segment, and uncertainty rather than smoothing everything into a single average.
2. Challenge
AI can act as a structured dissenter. It can identify assumptions that lack evidence, argue the strongest case against a preferred option, expose conflicts between departments, and ask what would have to be true for the strategy to work.
3. Simulate
AI can help teams explore how different buyers, competitors, partners, or employees may respond to a choice. Simulation does not prove what will happen. It helps the team discover questions, risks, and second-order effects worth validating.
4. Coordinate
AI can turn decisions into connected briefs, operating priorities, measures, and responsibilities. This reduces the gap between the strategy meeting and the daily work that must change because of it.
These capabilities are valuable precisely because they improve the quality and pace of human thinking. They become dangerous when the organization treats synthesis as truth, challenge as certainty, simulation as evidence, or coordination as commitment.
The buyer-intelligence gap in most AI strategy work
Many AI-assisted planning exercises begin with market reports, internal documents, financial data, and competitor websites. Those sources help explain the environment. They do not automatically explain the buyer’s decision.
A market may be growing while your offer remains irrelevant. A segment may fit your firmographic criteria while its buying group sees no urgency. A competitor may appear weaker on features while winning because buyers trust its implementation process. A strategic plan can be numerically coherent and still misunderstand what moves a buyer from interest to commitment.
This is where buyer intelligence changes the quality of the work. The team needs more than demographic and firmographic profiles. It needs maintained knowledge of pains, motivations, triggers, objections, decision criteria, internal pressures, stakeholder roles, alternatives, and the language buyers use to make sense of the decision.
Our internal-analysis framework helps connect capabilities to buyer value, while the guide to choosing a strategic planning framework helps select the right planning structure. AI adds leverage to both—but only when the buyer model beneath the analysis is deep enough to challenge generic conclusions.
This is also the operating-model argument in Andy Halko’s The Buyer-Centric Operating System: buyer understanding should not live inside a one-time workshop or a static persona. It should persist across the decisions made by marketing, sales, product, customer success, and leadership.
A buyer-centered AI strategic planning loop
The strongest AI-assisted planning process keeps evidence, interpretation, simulation, and accountability separate. That makes the work faster without making it careless.
-
Frame the decision
Start with the choice the strategy must inform. Are you entering a market, changing positioning, prioritizing a segment, building a capability, or reallocating resources? A vague prompt produces a vague plan.
-
Gather evidence with provenance
Bring together buyer research, performance data, competitive context, financial reality, team knowledge, and operating constraints. Preserve where each claim came from and when it was observed.
-
Expose assumptions
Ask AI to separate facts from interpretations and interpretations from guesses. Identify the assumptions that carry the greatest strategic risk and define what evidence would strengthen or disprove them.
-
Generate and simulate options
Create meaningfully different choices—not cosmetic variations of the same answer. Test each option against buyer needs, stakeholder objections, competitive responses, capability requirements, and plausible market shifts.
-
Make the tradeoff
Leadership chooses. Document why one option was selected, which alternatives were rejected, what the company is willing to stop doing, and who owns the consequences.
-
Learn and update
Define the signals that would confirm, challenge, or invalidate the strategy. Feed new buyer evidence and operating results back into the model instead of protecting the original plan from reality.
This separation echoes the NIST AI Risk Management Framework’s emphasis on mapping context before making go-or-no-go decisions. NIST’s broader Generative AI Profile also provides a useful governance lens for identifying, evaluating, and managing the distinct risks introduced by generative systems.
What weak AI strategy work looks like
AI does not create strategic discipline by itself. It can make undisciplined thinking look more complete.
| Weak pattern | What goes wrong | Better practice |
|---|---|---|
| Prompt first, evidence later | The model’s initial framing anchors the discussion before the team understands the situation | Form an independent view and assemble evidence before asking AI to challenge it |
| One answer presented as strategy | The team edits a recommendation instead of comparing real alternatives | Require meaningfully different options and state the tradeoffs in each |
| Generic buyer simulation | Outputs reflect stereotypes, surface personalization, or broad industry clichés | Ground simulation in maintained buyer intelligence and label unsupported inferences |
| Consensus by fluent prose | Confident language suppresses disagreement and makes uncertainty harder to see | Ask for contradictory evidence, minority interpretations, and conditions that would change the answer |
| No decision owner | The generated plan belongs to everyone and therefore no one | Name the person accountable for the choice, resources, measures, and learning loop |
Where AI creates the most strategic value
The 2026 Stanford AI Index reports that organizational adoption continues to rise rapidly, while AI use in strategy and corporate finance still trails many information-processing functions. That gap is understandable. Strategy has fewer repeatable answers, higher ambiguity, and more consequential tradeoffs.
AI creates the most value when it increases the breadth of evidence and the range of options without obscuring who must decide.
- Market entry: compare segments, surface adoption barriers, simulate stakeholder objections, and identify which assumptions require primary research.
- Positioning: pressure-test a promise against buyer pains, alternatives, proof requirements, and competitor claims.
- Product strategy: connect feature priorities to the decisions buyers and users are trying to make—not simply to request volume.
- Go-to-market planning: align marketing, sales, and customer success around the same buyer triggers, criteria, objections, and risks.
- Capability investment: determine whether an internal strength changes buyer value or merely improves internal efficiency.
The important pattern is consistent: AI expands the field of view. Buyer intelligence makes the view relevant. Leadership turns it into a choice.
Episode 20: lessons from building an AI planning platform
In this episode, Andy shares lessons from building Frictionless, an early platform designed to support structured strategic thinking, progressive exercises, and more complete planning workflows. The product has evolved, but the central question remains current: how do we use AI to improve strategic work without confusing generated output with judgment?
Do not outsource the part that makes strategy matter
AI can help a team move beyond blank pages, incomplete research, and narrow thinking. It can expose contradictions, accelerate analysis, generate alternatives, and keep decisions connected to execution.
But strategy still requires a point of view.
Someone must decide which buyer matters most, which problem deserves priority, which advantage is worth building, which opportunity should be declined, and what the organization will do when evidence challenges the plan.
Use AI to make those choices harder to fake and easier to examine. Do not use it to avoid making them.