AI Product Strategy Memo Template (6-Page Amazon-Style)
TL;DR
Slide decks hide weak thinking. Amazon-style 6-page narrative memos surface it. This template adapts the format for AI product strategy: customer story, market context, product vision, AI approach, plan, asks. The format forces clarity; the discipline of writing it makes the strategy survive reading aloud in a quiet room.
Why Narrative Memos Beat Decks for Strategy
Decks let weak thinking hide between bullet points. Narrative memos force you to write "because" and "therefore" — to make the logic visible. For AI product strategy, where a single capability assumption can collapse the whole plan, the discipline of narrative is exactly what's needed.
Page 1: Customer story
The user, the pain, the moment of need. Concrete, specific, evidenced. The whole memo earns its space here.
Page 2: Market context
Why now. What changed. Who's competing. What customers expect in 2026.
Page 3: Product vision
The 18-month picture. Specific features, specific surfaces, specific outcomes.
Page 4: AI approach
Capability claim, model choice, eval methodology, cost model. The differentiator section.
Page 5: Plan and risks
Phased plan, kill criteria, top 3 risks with mitigations.
Page 6: Asks and decisions
What you're asking the audience to decide today. Specific resources, deadlines, alternatives.
Page 1 — Customer Story
The customer story is the foundation. If page 1 doesn't convince readers there's a real, specific user with a real, specific pain, nothing in pages 2-6 matters. This is also the page that's hardest to write — most teams skip it or fake it.
Named user persona
Specific role, specific company size, specific workflow. "The customer" is too generic.
Specific moment of pain
When the pain hits, what they're doing, what they wish they could do instead. A concrete vignette.
Quantified impact
Time wasted, money lost, opportunities missed. With evidence — research, support tickets, or behavior data.
Quote
One real quote from a real user. Anchors the story in voice. Replaces 100 words of summary.
What customers do today
The current workaround. Reveals how badly they need help — and what they've already tried.
Page 4 — AI Approach (The Differentiator)
Page 4 is where AI strategy memos are won or lost. Generic AI talk loses; specific capability claims tied to evidence win. This is also the page that benefits most from technical reviewers; have an engineer read it before you ship.
Capability claim with evidence
"The model can do X with Y reliability — verified in our 100-case eval." Specific. Sourced.
Why now (capability shift)
What changed in models or costs that makes this feasible today. Anchors the timing argument.
Architecture sketch
Model choice, retrieval if any, eval methodology, fallback plan. Half a page max; reference appendix for detail.
Cost-at-scale model
Per-request × volume × peak factor. Margins at projected scale. The CFO line.
Write Strategy Memos That Land
The AI PM Masterclass walks through real strategy memos with cuts, rewrites, and exec-style read-aloud reviews — taught by a Salesforce Sr. Director PM.
Running the Read-Aloud Meeting
First 30 minutes: silent reading
Everyone reads the memo in the room, in silence. Phones away. The discomfort is the point — it forces engagement.
Next 60 minutes: discussion
Open questions, challenges, debate. The memo is the artifact under critique; the author is in the room to answer, not defend.
Last 15 minutes: decisions
What does the room agree on? What's open? What's next? Document specifically.
Post-meeting: revised memo
Author updates the memo based on the discussion. New version becomes the source of truth for the next read-aloud or the eventual approval.
Memo Anti-Patterns
Padding to hit 6 pages
If you don't have 6 pages of substance, write fewer. Padding is worse than brevity. Cut to actual signal.
Bullet-point disguised as prose
Sentences with 3 commas pretending to be paragraphs. Real prose explains why; bullets list what.
AI mentioned only in passing
If page 4 doesn't name model, eval, and cost, you don't have an AI strategy. Be specific.
No real risks listed
"Execution risk" doesn't count. List what could actually break this — capability gap, vendor change, customer adoption.
Vague asks
"Looking forward to feedback" is not an ask. Specific resources, deadlines, and decision points only.