TL;DR
You have 4 versions of your product story scattered across the website, the deck, the App Store, and the founder’s podcast voice. AI can hold April Dunford’s five-component structure steady across a dozen drafts and stress-test whether a sentence could also describe a competitor. It cannot pick the right category for you — that is a strategic call. Feed it competitive alternatives, real customer quotes, and proof; ask for 5 candidates plus stress tests; then validate by reading the winner to 3 actual customers and watching for the nod. Use a reasoning model (Claude Opus 4.7 or GPT-5.5 Thinking) for the strategy pass, a cheaper one for cleanup.
The task
Your team has four descriptions of the product: one on the website hero, one in the sales deck, one in the App Store listing, one your founder ad-libs on podcasts. They are close enough to feel coordinated but different enough that ads, content, and sales-call openers don’t reinforce each other. You want a single positioning statement everyone aligns on — category, buyer, and what you displace — and you want 5 candidates drafted by lunch so leadership can argue over them in the afternoon.
Get the framework right first
Most “positioning statement” templates (the “For [X] who [Y], [product] is a [category] that [benefit]” Madlib) are the trap. As April Dunford puts it bluntly: using a fill-in-the-blank statement to do positioning “is stupid… how do we know what the answers are for the blanks? We’re supposed to just make it up?” The statement is the documentation; the thinking comes first.
That thinking has five components, and the order matters because each one feeds the next (April Dunford, Obviously Awesome):
| # | Component | Plain-English question |
|---|---|---|
| 1 | Competitive alternatives | If you didn’t exist, what would your best customers use instead? (Often a spreadsheet, an intern, or nothing — not a named rival.) |
| 2 | Unique attributes | What do you have that those alternatives don’t? |
| 3 | Differentiated value | What do those attributes let the customer do or avoid? |
| 4 | Best-fit customers | Who cares about that value the most, and why? |
| 5 | Market category | What context makes the value obvious to that customer? |
The old version of this article (and most blog posts) starts at category. Dunford ends there — you earn the right to name the category only after you know the alternatives and the value. AI is happy to start anywhere, so you have to feed it in this order.
Where AI helps — and where it does not
AI is good at applying that five-component structure consistently across many drafts, and at stress-testing whether a candidate would also describe a competitor word-for-word.
What AI cannot do: pick the right category to anchor in. That decision involves where the budget already lives, who you’re recruiting buyers away from, and how the largest player is shaping the market. AI also can’t tell you whether your “unique value” is true — it has no way to know your retention curve beats the alternative’s. Test the output two ways: read it to 3 actual customers and watch for the nod, then read it to a competitor’s salesperson and see if they could plausibly steal it.
A common failure mode: AI writes positioning that sounds elegant but could describe Notion, Linear, and Asana equally well. The fix is to force category specificity and ban “best-in-class,” “innovative,” and “leading” — Dunford flags these as hollow claims that establish no differentiation.
What to feed the AI
- The competitive alternative — the tool or behavior they default to when you don’t exist (a spreadsheet, a Slack channel, an intern), not the obvious named rival
- The 1-2 things you genuinely do better than that alternative, with one piece of evidence each (a retention number, a customer quote, a feature competitors lack)
- The differentiated value those attributes unlock (what the customer gets to do or stop doing)
- The best-fit customer plus why they care more than anyone else
- The “why now” — what shifted in the market that opened this gap
- 3-5 customer quotes describing the product in their own words (these usually reveal the real category)
- Your team’s current top 3 candidate statements (so AI can diff and improve)
- Anti-examples — competitor positioning you don’t want to mirror
Which model to use
As of June 2026, treat this as a two-pass job:
| Pass | Model | Why |
|---|---|---|
| Strategy + stress tests | Claude Opus 4.7 or GPT-5.5 (Thinking) | Reasoning models hold the 5-component logic and catch “a competitor could say this too” |
| Tighten / variants / hero line | Sonnet 4.6 or GPT-5.5 (Instant) | Cheaper, fast, fine once the strategy is locked |
All of these read a full quote dump without trouble — Opus 4.7, Sonnet 4.6, and Gemini 3.1 Pro carry a 1M-token context window, and even ChatGPT Plus handles roughly 320 pages in-app. Paste your raw call notes; don’t pre-summarize them, because the literal customer phrasing is the most valuable input.
Copy-ready prompt
Write a product positioning statement using April Dunford's five-component method.
Work the components in THIS order; do not start at category:
1. Competitive alternative: {tool or behavior they default to today}
2. Unique attributes: {1-2 things you have that the alternative lacks}
3. Differentiated value, with evidence: {what those attributes let the customer do/avoid + proof}
4. Best-fit customer: {who, in 5 words including stage and constraint, and why they care most}
5. Market category: {the frame of reference — only after the above}
Why now: {market shift}
Customer quotes (their words): {3-5 quotes}
Competitor positioning to avoid mirroring: {paste}
Return:
1) Positioning statement: one paragraph, 4 sentences max, plain English. No "best-in-class", "innovative", "leading".
2) 3 variants — same core, different emphasis (one buyer-first, one alternative-first, one why-now-first).
3) Three stress tests with verdicts:
a) Would the buyer say "yes, that's me"? Quote the customer line that confirms it.
b) Could a top competitor truthfully claim the same sentence? If yes, rewrite until only you can.
c) Could the whole team explain it the same way in 2 weeks?
4) The specific clichés this statement avoids.
5) Three questions to ask the next 5 customers to validate it.Shorter variant — sharpen existing positioning
Below is our current positioning paragraph and 5 customer quotes. Identify the 2 weakest words/phrases (vague, claimable by competitors, jargon) and rewrite each with a sharper alternative grounded in the quotes. Show the before/after diff.
Current: {paste}
Quotes: {paste 5}Sample output
Strong positioning: “For solo founders who want to ship faster than they raise (alternative today: hiring a part-time dev), [Product] is the AI co-builder that pairs with you inside a real IDE, not a chat-only assistant. Unlike chat tools, it replays your last 30 days of project decisions on demand, so you never re-explain context every morning. We exist now because the unit economics of solo building finally make a code co-pilot a substitute for a junior hire, not just an editor plugin.”
Strong stress-test verdict: “Test (b) FAILS as written — ‘remembers your project’ is also claimable by [Competitor X]. Rewrite: ‘replays your last 30 days of decisions on demand’ — that is a feature we have and they don’t.”
Note how the rewrite moved from a value claim anyone can make to a specific attribute (component 2) that proves the value (component 3). That is the whole game.
How to refine
- Force category specificity: “Rewrite the category in 7 words including the buyer’s stage. ‘SaaS’ fails; ‘design ops platform for in-house brand teams’ passes.”
- Name the actual alternative: “The alternative is the behavior they default to today, not the obvious competitor. It could be a spreadsheet, an intern, or a Slack channel. What do they use when you don’t exist?”
- Pressure-test uniqueness: “Read the second sentence aloud as if you were [Competitor X]. If they could say it without lying, rewrite until they can’t.”
- Anchor in customer language: “Replace any abstract noun phrase with a quoted phrase from the customer quotes.”
- Demand the why-now: “Add one sentence on why this positioning is winnable in 2026 specifically, not 2024. If the answer is ‘it’s not time-sensitive,’ the positioning isn’t sharp enough.”
For physical-product teams, the positioning statement also has to survive on a shelf. Pair this with the 10-minute packaging concept brainstorm so the visual and verbal stories land the same way. To get sharper customer quotes feeding step 6, run customer discovery interviews first.
Common mistakes
- Positioning as “the best [category]”. Every competitor says this and buyers tune it out
- Targeting “everyone”. You target no one, and your marketing copy goes generic
- Skipping the alternative. Buyers always compare; if you don’t name the comparison, they pick the wrong reference
- Drafting in a vacuum. Without 3-5 customer quotes, you’re guessing at the category your buyer actually puts you in
- Starting at category. Dunford’s method ends there — name it only after you know the alternatives and value
- Confusing positioning with messaging. Positioning is the internal anchor; messaging is the external phrasing per channel
- Updating positioning to chase last quarter’s deals. Chase the structural buyer, not the noisy one
- Letting marketing own positioning alone. Product, sales, and CS each see a different distortion; align them or your positioning fragments again in 6 weeks
- Writing positioning as a paragraph but never testing the one-sentence website hero version derived from it
FAQ
- Is a positioning statement the same as a value proposition?: No. Positioning is the strategic decision about which market context and alternatives you compete against; the value proposition and all your messaging are outputs of it. Positioning comes first and feeds messaging, not the other way around.
- How long should the statement be?: One paragraph internally (4 sentences max). One sentence on the website hero. They share the same anchor; the hero is the 7-word reduction of the paragraph.
- How often should I update it?: When the buyer shifts (you crossed from SMB to mid-market), the competitive landscape shifts (a giant entered the category), or a 50-quote customer dataset contradicts the current positioning. Not whenever leadership wants to refresh the deck.
- Should I A/B test positioning copy?: Test the headlines derived from positioning, not the positioning itself. The positioning is the strategic anchor; headlines are the executions.
- What if my buyer disagrees with the category I picked?: That’s the signal to change the category or change the buyer. The buyer’s mental model wins; positioning that fights it loses.
- How do I know it’s working?: Sales discovery calls open with “yeah, we were looking for an X that does Y.” Those are your category and buyer words coming back. Inbound demo requests start naming the alternative explicitly.