Product Strategy Sparring with AI: Beyond the Blank Template

Why a strategy template doesn't produce a strategy

Most "strategy docs" are templates filled with aspirational language that could describe almost any product: become the market leader, delight customers, grow sustainably. None of that is false. None of it is a decision either. A real strategy names what you're explicitly choosing not to do, and choosing not to do something is uncomfortable — it closes a door, it disappoints a stakeholder who wanted the other thing. A template doesn't force that discomfort. It just asks you to fill in blanks, and blanks get filled with whatever sounds safe.

An AI chatbot prompted for "a product strategy" makes this worse, not better — it's exceptionally good at generating fluent, confident-sounding text that avoids committing to anything specific, because vague is easier to generate than decisive. The fix isn't a better strategy prompt template. It's changing what you're asking the agent to do: not write a document, but argue with you until a real position survives.

A strategy you can't defend under pushback isn't a strategy — it's a mood board with bullet points.

This matters more than it sounds like a nitpick, because the cost of a vague strategy isn't paid immediately — it's paid three months later, when a team ships something reasonable-sounding that turns out to be aimed at nothing in particular, and nobody can point to the document and say "this is why we didn't do that instead." A template that never forced the trade-off never gave anyone that ability. Sparring is slower up front and cheaper everywhere after.

Step 0 — Ground the agent in your actual constraints, not industry commentary

Sparring only sharpens thinking if the pushback is grounded in your real situation — your team size, your runway, what's already been tried and failed, your actual competitive position. An agent with no context defaults to generic strategy commentary borrowed from wherever it's seen "product strategy" discussed, which produces pushback that sounds smart but doesn't apply to you. Before sparring, make sure your workspace has:

• context/company.md — stage, resourcing, what leadership has already ruled in or out
• context/product.md — what exists today, what's been tried, what's currently working and not
• context/users.md — your actual segments, so "differentiation" isn't argued in the abstract

This is the same discipline behind every artifact in an agentic PM workflow: the agent is only as sharp as the reality you've given it to push against.

Step 1 — Ask for a diagnosis, not a direction

The instinct is to bring a solution ("we should focus on enterprise") and ask the agent to build a strategy around it. Resist that — you've skipped the actual strategic work. The toolkit's /product-strategy command uses Reforge's 4-type model precisely to force this ordering: diagnose what kind of strategic situation you're actually in before anyone proposes a direction. Here's a starting prompt that does the same thing without the toolkit:

Read context/company.md, context/product.md and context/users.md first.

I want to spar on a product strategy question, not receive a
finished document. Do this in order:

1. Ask me clarifying questions about the market, our constraints,
   and what's already been tried, until you have enough to reason
   from — don't skip to advice.
2. Diagnose what TYPE of strategic problem this actually is before
   suggesting any direction (e.g. a market-timing question, a
   differentiation question, a business-model question, or a
   go-to-market sequencing question). State which one, and why,
   grounded in context/company.md — not generic industry patterns.
3. Push back on my first instinct with the strongest real
   counter-argument you can make from our actual constraints.
4. Do NOT hand me a strategy statement yet. Force a trade-off:
   what would we have to explicitly stop doing or stop chasing
   to commit to this?

Strategic question: [describe the decision you're facing]

Step 2 is what separates this from a chat session — misdiagnosing the type of problem is the single most common strategy failure, and it's invisible until someone tests the plan against a situation it was never actually suited for. A market-timing problem solved like a differentiation problem produces a strategy that's internally coherent and pointed at the wrong target entirely.

Step 2 — Pressure-test the narrative like an exec would, before an exec does

Once a direction survives the diagnosis, don't stop — stress-test it the way a skeptical VP actually would in the room: why now and not six months ago, why this and not the obvious alternative, what happens if a competitor moves first, what's the failure mode if the core assumption is wrong. A structured pushback pass (the toolkit runs this as part of /product-strategy, using the Racecar framework to keep the pressure-testing consistent) catches the gaps in your reasoning while the cost of being wrong is a slightly awkward agent conversation, not a credibility hit in front of leadership.

Ask directly: "What's the strongest argument a skeptical exec would make against this, and how would I actually answer it — not a talking point, an answer that survives a follow-up question?" If the agent can't construct a strong counter-argument, either the strategy is genuinely solid, or you haven't given it enough real constraints to argue with. Assume the second one first.

Run this more than once, and vary the angle each pass: one round on market timing, one on resourcing reality, one on what a competitor's most aggressive response would look like. A strategy that survives three different lines of attack is meaningfully more trustworthy than one that survived a single generic "any concerns?" pass — and each round takes minutes, not the days a real cross-functional review cycle would cost you if the gap surfaced there instead.

Step 3 — From sparred strategy to a bet you can frame

A strategy that survived sparring is still not an artifact anyone can act on — it's a direction. The next step is framing it as a concrete bet: an 8-section initiative document, built for cross-functional alignment, that a team can actually plan against. Chaining the two in the same grounded session means the initiative inherits the reasoning instead of restating it from scratch — including the PRD downstream of it, once a specific solution is chosen. This is the same handoff discipline covered in the full command list: strategy feeds the initiative, the initiative feeds delivery.

Keep the reasoning attached even after the strategy ships as a slide. Six months from now, when someone asks "didn't we consider the other option," the sparring transcript is the answer — not a reconstructed memory of a meeting nobody took clean notes in. That's a small thing until the day it saves you from re-litigating a decision you already made carefully the first time.

Common mistakes to avoid

• Bringing a solution and calling it a strategy. "Focus on enterprise" is a direction, not a diagnosis. Ask what problem type you're actually solving first.
• Accepting the first draft. A strategy that hasn't been argued with hasn't been tested. Push for the counter-argument before you trust the conclusion.
• Sparring without real constraints in context. Ungrounded pushback sounds smart and applies to nobody — it's industry commentary wearing a strategy costume.
• Avoiding the trade-off question. If nothing was explicitly ruled out, you don't have a strategy yet, no matter how polished the document reads.
• Treating strategy and initiative as the same artifact. Strategy is the direction; the initiative is the specific, plannable bet. Collapsing them skips the diagnosis.

FAQ