AI is revolutionizing strategic decision-making

Paper highlights

The tools of strategy were designed around limited human attention, memory, and processing capacity. AI does not expand the biological limits of the people in the room. It expands the capacity of the system making strategy: people, LLMs, data, workflows, and decision rules working together.

That system can examine more alternatives, maintain richer models, and challenge proposals more systematically than a conventional planning process. We call this an unbounding of rationality—and, in practical terms, an unbounding of strategy. People still frame the question, inspect evidence, resolve trade-offs, and remain accountable.

Three changes to the strategy process

• Search: generate, filter, cluster, and rank a much larger option set before leaders narrow it.
• Representation: replace periodic, simplified maps and forecasts with richer models that update as evidence changes.
• Aggregation: use creator–critic workflows, simulated stakeholders, and adversarial agents to pressure-test proposals and surface objections that hierarchy or groupthink may suppress.

These are changes in kind, not only speed. Documented applications span acquisition search, small-business credit, and AI-supported product development.

Where advantage comes from

Access to the same LLM will not distinguish a firm for long. Advantage can persist in proprietary data, hard-to-copy processes, faster learning, feedback loops, and embedded routines.

A leader’s playbook

• Require an AI-expanded list of alternatives considered and rejected before major choices.
• Turn one important static artifact into a living representation that updates with new evidence.
• Ask separate AI roles to create, criticize, and defend a proposal, then have people examine the disagreement.
• Make the process standard where it solves a recurring problem; an optional chatbot at the edge of work will not change how the organization thinks.

AI output can still be hallucinated, inconsistent, or irrelevant. The hybrid strategist therefore acts as an architect: defining objectives, designing the workflow, demanding evidence and critique, and deciding where machine analysis yields to human judgment.

How to use this article

Cite this for

• A practical formulation of human–AI strategy systems that expand search, representation, and aggregation.
• The idea of the hybrid strategist as an architect of decision workflows.
• The argument that advantage comes from proprietary data, embedded processes, feedback loops, and learning rather than generic model access alone.

Useful for teaching

• How AI changes the strategy process rather than merely speeding up existing analysis.
• Why an optional chatbot is different from redesigning a recurring decision workflow.
• How leaders can assign AI roles for generation, critique, defense, and evidence checking.

Careful claim

The article argues that AI can expand the capacity of designed strategy systems; it is not a claim that standalone AI automatically creates advantage or replaces human accountability.

Abstract

This forthcoming Harvard Business Review article argues that AI changes strategic decision-making by expanding three cognitive tasks at its core: search, representation, and aggregation. AI can generate and screen far more alternatives than human teams, replace static strategy frameworks with richer and more dynamic models of markets and competitors, and enable structured forms of challenge and deliberation that are less constrained by hierarchy, fatigue, and time pressure.

The article then addresses a central strategic objection to AI adoption: if many firms can access similar models, where does advantage come from? Its answer is that durable advantage depends less on generic model access than on the proprietary data, embedded processes, and speed of organizational adaptation that firms build around those models. The piece closes with a practical playbook for leaders who want to redesign strategic decision-making and shift the strategist’s role from analyst to architect of an AI-augmented process.