AI Can Write Strategy — But It Can’t Be a Strategist
On a Tuesday morning, Arjun Velasco sat in a glass-walled conference room, scrolling through a document that had taken less than a minute to produce. He had asked an AI system to generate a strategy for a new outreach initiative. The result was polished, organized, and oddly reassuring.
Across from him, Lila Moreno read the same document in silence. She had spent years navigating complex institutional projects — some successful, many not. Her experience had made her wary of anything that looked too complete too quickly.
“It’s impressive,” Arjun said. “It’s already structured better than what we usually start with.”
Lila nodded slightly. “Yes,” she said. “It’s well written. But I’m not sure it’s a strategy.”
The Seduction of Fluent Strategy
This is where AI is most persuasive. It produces language that feels authoritative. The sections are in the right order. The tone is measured. The logic appears sound.
Fluency carries weight. When something reads clearly, we tend to assume the thinking behind it is equally clear. The presentation creates a sense of progress, even when the underlying issues remain unresolved.
Arjun found himself drawn to that clarity. The document reduced ambiguity. It suggested that the problem was already well understood, that the path forward was largely mapped.
Lila had seen this dynamic before. Not with AI, but in carefully prepared plans that created the appearance of alignment before real agreement had been reached. The risk was subtle: the work felt further along than it actually was.
What Strategy Actually Requires
Strategy is not primarily about describing possibilities. It is about choosing among them.
Those choices are rarely comfortable. They involve trade-offs, constraints, and incomplete information. To commit to one direction is to exclude others.
The document in front of them outlined several strong options. Each was plausible. Each had its own rationale. None was clearly prioritized.
Arjun pointed to a section labeled “Recommended Approach.” “This seems like a good direction,” he said.
Lila looked at it briefly. “Is it the direction?” she asked.
He paused. The document had not required him to decide. It had allowed him to remain in a space of well-articulated options.
Patterns Without Context
The AI had drawn on a wide range of examples — campaigns, initiatives, and frameworks that had worked elsewhere. It recombined them into something that resembled a tailored plan.
This is a real capability. AI systems are exceptionally good at identifying patterns across domains and presenting them in coherent form.
But strategy depends on context. Arjun’s organization had its own internal dynamics: competing priorities, informal influence, and a history that shaped how new initiatives were received.
None of that appeared in the document.
Lila tapped a paragraph outlining stakeholder engagement. “This assumes cooperation,” she said. “Do we have that?”
Arjun did not answer immediately.
When Precedent Stops Helping
Much of strategic thinking begins with analogy. We look for similar situations and adapt what worked before. AI excels at this kind of reasoning.
The difficulty comes when the situation diverges from precedent. When conditions shift, familiar patterns can become misleading.
Arjun’s initiative involved a new audience and a compressed timeline. The AI-generated plan reflected stable, well-resourced contexts. It did not account for the specific constraints they were facing.
Lila leaned back. “This feels like it’s built on examples where things behaved predictably,” she said.
“And here?” Arjun asked.
“Here,” she replied, “we’re not sure they will.”
The Absence of Consequence
There was another difference, less visible but more fundamental.
If this initiative failed, Arjun would be accountable. His credibility would be affected. Future opportunities might depend on the outcome.
The AI system carried no such burden. It did not experience risk. It did not prioritize under pressure. It would never have to defend its recommendations in a tense board meeting, or live with the consequences of a poorly timed commitment.
This is not a limitation that better prompting can fix. It is structural. An AI that faces no consequences cannot model what it feels like to face them. And strategy — real strategy — is inseparable from that weight. The person who chooses knows what is at stake. That knowledge shapes which options feel viable, which risks feel acceptable, and which commitments feel possible.
Get JM Bonthous’s stories in your inbox
Join Medium for free to get updates from this writer.
As a result, the AI’s recommendations were balanced and careful. It presented options, acknowledged uncertainties, and avoided strong commitments. That made the document feel thoughtful. It also made it non-committal.
“At some point,” Lila said, “you have to decide what you’re willing to risk.”
The document did not answer that question.
The Role of Power
Strategy is shaped not only by logic but by power.
Organizations operate through formal roles and informal influence. Decisions are often determined by individuals who may not appear in official structures.
The AI-generated plan described a clean process: defined roles, orderly approvals, clear progression from concept to execution.
Arjun knew the reality would be less straightforward. Certain stakeholders would need to be persuaded. Others might resist quietly. A few would have decisive influence regardless of their formal position.
Lila asked, “Who actually has the authority to stop this?”
Arjun named a senior figure who was not mentioned in the plan.
“Then that’s where your strategy starts,” she said.
The Tool That Never Pushes Back
AI systems are designed to be helpful. They respond to prompts by developing and reinforcing the direction implied in the question.
Arjun had noticed this. When he emphasized a particular idea, the system expanded on it. When he shifted focus, it adapted smoothly. This responsiveness can be useful. It supports exploration and speeds up early thinking.
But it also reduces friction. The system rarely challenges assumptions unless explicitly asked to do so — and even then, the challenge arrives in measured, diplomatic language. It does not have a stake in the argument. It does not experience the discomfort of telling someone their preferred approach is built on a flawed premise.
Lila’s role in the conversation was different. She introduced tension deliberately. She questioned the premises, not just the conclusions.
“Who is going to tell you this won’t work?” she asked.
Arjun smiled. “You seem to be doing that.”
“That’s part of the process,” she said.
What AI Contributes
Despite her reservations, Lila did not dismiss the document.
“It’s a strong starting point,” she said. “It’s just not the end point.”
Arjun had used the AI to generate structure, explore options, and articulate ideas more clearly. It had accelerated the early stages of thinking.
In that sense, it was valuable. It reduced the time needed to move from vague concepts to something more concrete.
“It helps you see the landscape,” Lila said. “But you still have to choose a path.”
Knowing and Deciding
By the end of the meeting, the document looked different. Arjun had marked sections to keep, others to revise, and several to question.
The AI had provided a map of possibilities. It had not determined a course of action.
Strategy exists in the space between analysis and commitment. It involves choosing under conditions where certainty is not available and consequences are real.
AI can describe strategies with remarkable clarity. It can assemble patterns, organize ideas, and simulate reasoning. But it does not share the conditions under which strategy is made — the uncertainty, the competing pressures, the awareness that something important is on the line.
What Arjun left the room with was not a strategy. It was material he could now shape into one. The document had done its job. The rest was his.
Jean Marie Bonthous (publishing as JM Bonthous) is the author of more than two dozen books, including six on the human side of AI, six about filmmaking, and four about digital/AI art. See his latest books: www.jmbonthous.com
He writes three blogs on Medium:
About the human dimensions of AI: AI in Real Life
About AI art: The Algorithmic Eye
About AI filmmaking: The Solitary Frame
