Force a strategy through several conflicting perspectives on purpose so you hear the objections before the room does.
Most strategy reviews collapse into one house view with a polite dissent slide. Conflicting perspectives are the point. You want the CFO reading, the general counsel reading, the enterprise buyer reading, and the competitor reading to collide while you still control the clock.
More than two sides
Argue both sides is the right tool when the choice is binary and you need the strongest case for each option. Multi-perspective review is different. It maps the leadership seating chart: who will speak, what they optimize for, and which facts they will treat as decisive. A pricing change may look clean to product and fatal to finance. A partnership may look strategic to business development and risky to security. A market entry may excite sales and terrify operations. Hearing those seats in private is cheaper than hearing them in the offsite.
Pick three to six seats that will actually matter for this decision. More than six usually repeats the same conflict. Fewer than three usually collapses into a simple for-and-against. Name the people or functions who will be in the real room when you can; keep labels generic when the document may travel.
Build the perspective pack
Write each perspective as a short brief, not a persona cartoon. Include what that seat measures, what evidence they trust, what they have rejected before, and the question they will ask first. Attach the same source pack to every seat so the conflict is about interpretation and priorities, not about secret data. If one seat needs a document the others should not see, run that seat as a separate pass with clear redaction rules.
Run the collision
Ask the review to answer, for each perspective: what would make this strategy unacceptable, what evidence is missing, and what change would make the seat support the move. Then ask for the conflicts between seats that the current draft papers over. Pingpong's sequential multi-model review helps here because later models can score earlier answers from a new seat instead of restating the same summary. Keep a running list of unresolved conflicts. Those become agenda items for the real meeting, or reasons to change the strategy before you socialize it.
Stop when you have a short conflict map: which seats agree, which seats block, and which facts would move a blocker. If every seat somehow agrees, your perspective briefs were too similar. Rewrite the seats until at least one material conflict appears, then decide whether to change the strategy or walk into the room ready to answer that conflict.
Continue with war-gaming the strategy, high-stakes AI review, find holes in your strategy, and challenge your strategy. Process links: first review, how it works.