TL;DR
Feed an AI model your product, your 3-5 real competitive alternatives, and 2-3 audience segments, then have it return 3 genuinely different positioning options — each built on April Dunford’s five components (alternatives, unique attributes, value, target segment, market category) and each with an explicit “who this is NOT for.” Use a strong reasoning model (Claude Opus 4.7 or GPT-5.5) for the strategy pass. Then validate with real buyers — a preference test on a panel like Wynter returns answers in 12-48 hours — before you commit. AI gets you three credible drafts; customer language picks the winner.
The task
You are repositioning an existing product, or launching a new one, and you cannot decide how to frame it. You know the audience and the competitors, but you have no clean articulation. You want 3 genuinely different positioning candidates to test — not three rephrasings of the same idea — each with a tagline, a real differentiator, and an explicit “who this is not for.”
The trap most teams fall into: they ask AI for “positioning” with a thin product blurb and get back three flavors of “the powerful, easy way to do X.” Three taglines, one idea. The fix is to make the model reason through a real framework and force the options apart on audience, value lens, and named enemy.
When AI is the right tool
- You have a product description, a sense of the audience, and at least 3 competitors named.
- You have run yourself in circles writing your own positioning.
- You want to compare distinct directions before committing.
- You will validate the options with real conversations, not just internal debate.
When not to rely on AI alone
- For category-creating products. Positioning a new category is a multi-month exercise grounded in user research; AI gives raw material, not the answer.
- For regulated industries where claims must clear legal review.
- When founder conviction matters more than the data. Sometimes the right positioning is the unintuitive one only you can see. AI averages toward defensible, not bold.
Pick the model
This is a reasoning and judgment task, not a copy task, so use a top-tier model rather than a fast one. As of June 2026:
| Model | Plan to run it on | Why for positioning |
|---|---|---|
| Claude Opus 4.7 | Claude Pro $20/mo or Max $100+/mo | Strongest at nuanced strategy reasoning and resisting generic copy; 1M-token context fits a full competitor dossier |
| GPT-5.5 (Thinking) | ChatGPT Plus $20/mo | Sharp, fast structured output; good second opinion to compare against Opus |
| Gemini 3.1 Pro | Google AI Pro $19.99/mo | 1M context; useful if your inputs already live in Google Docs/Workspace |
Run the same prompt on two models and compare. The disagreements are often where the interesting positioning hides. Avoid free-tier or “instant” modes for this — they swerve toward the safe, generic answer that is exactly the failure mode you are trying to escape.
What to feed the AI
Positioning quality is capped by input quality. Bring:
- Product: what it is, what it does, who uses it today.
- Competitive alternatives (April Dunford’s first lever): not just named rivals but what a buyer would do instead — a spreadsheet, a manual process, a freelancer, or nothing. List 3-5 with their tagline and a one-line read on how they position.
- Audience: 2-3 segments with their pain, what they use now, and the trigger that would make them switch.
- Constraints: brand voice, price tier, channel, things you will not say.
- Status quo positioning, if any.
Copy-ready prompt
This prompt bakes April Dunford’s five positioning components into the structure so the model reasons like a positioning consultant instead of a copywriter. Replace each [bracketed placeholder] with your own input.
You are a B2B product positioning strategist using April Dunford's
five-component method (competitive alternatives, unique attributes,
value, target segment, market category).
Propose 3 DISTINCT positioning options for our product.
Product: [what it is, what it does, who uses it today]
Competitive alternatives (what a buyer would do instead of us):
- [alternative 1: tagline + 1 line on how it is positioned]
- [alternative 2: ...]
- [alternative 3: ...]
Audience segments:
- [segment 1: pain, current tool, switching trigger]
- [segment 2: ...]
Constraints:
- Brand voice: [3-5 adjectives]
- Things we will NOT say: [list]
- Price tier: [budget / mid / premium]
For EACH of 3 positioning options, return:
1. A 3-word name in the form "the X for Y".
2. A one-line tagline.
3. Market category we are framing the product into.
4. Key differentiator: a unique attribute competitors structurally
cannot or will not match (not a feature claim).
5. Value: the payoff of that attribute for the target buyer.
6. Who it is FOR: specific segment + their switching trigger.
7. Who it is NOT for: the customer this positioning loses.
8. Risk: the single strongest objection a skeptic will raise.
Make the 3 options genuinely different — different audience,
different value lens, and a different named "enemy" each. After the
three, add one sentence on which option you would test first and why.Recommended output structure
A useful positioning artifact has 3 truly different options. Each option names a market category, a differentiator the model can defend as structural (not a feature), a specific “for whom + when,” a specific “not for whom” (the discipline most teams skip), and a clear risk to weigh. If you get three options with the same enemy and the same audience, the inputs were too thin — add more on competitive alternatives and switching triggers, then rerun.
How to check the output
- Are the 3 options actually different? Test by asking: would each one lose a different competitor head-to-head? If they all fight the same rival, you have one option in three outfits.
- Does each have a clear “not for”? If everyone is the target, positioning has failed.
- Is the differentiator real? It should name something competitors structurally cannot match — a business model, a data moat, a focus they cannot serve — not a feature they could copy next quarter.
- Does the category make sense to a cold buyer? If a stranger cannot place your product in a known mental shelf in five seconds, the category framing is fighting you.
Common mistakes
- Trying to be for everyone. “Powerful yet easy” is a non-statement.
- No defined enemy. Strong positioning has a clear “instead of X.”
- Choosing the safest option in the room. Safe positioning means a crowded category fight you are not set up to win.
- Picking before testing. Take each candidate to real customers; internal debate cannot tell you which line makes a buyer lean in.
- Treating AI output as final. It is a strong first draft. The job is to validate, then rewrite in your customers’ actual words.
Next steps: validate before you commit
Pick the 2 strongest options and write a one-paragraph pitch for each. Then test them against real buyers — do not let the loudest voice in the room decide:
- Customer conversations. Run 5 thirty-minute conversations per option (10 total). Score each on one question: who earned an unprompted “I would pay for that today”?
- Or a structured preference test. Tools like Wynter show your target ICP two or three versions and ask which they prefer and why, returning results from real B2B buyers in 12-48 hours — faster than scheduling ten calls.
- Refine with the verbatim. Feed the exact phrases customers used back into the model and ask it to rewrite the winning positioning in that language. The words your buyers already use convert better than anything AI invents.
FAQ
Which AI model is best for positioning work? A top reasoning model: Claude Opus 4.7 (via Claude Pro at $20/mo) or GPT-5.5 in Thinking mode (ChatGPT Plus at $20/mo), as of June 2026. Avoid free or “instant” tiers — they default to the generic, safe answer that is exactly what good positioning has to avoid.
Should the AI also generate the website headline? Not in this pass. Choose the positioning first, then feed the winner back and ask for 10 headline variants. Mixing the two muddies both.
How many positioning options should I test? Start with 3 candidates, collapse to 2 after team review, and choose 1 only after customer validation. Three is enough to force real difference without diluting the test.
How often should I revisit positioning? At least annually, or whenever a major competitor shifts its messaging, you move up- or down-market, or a new category emerges around you.
How do I move from “3 options” to one approved internal statement? Use AI positioning statement — it collapses the options into a single committable sentence with the assumptions named.
What if I’m repositioning against a specific competitor’s content? Reverse-engineer their content first with the AI competitor content teardown, so the new positioning answers their actual messaging, not just their tagline.
Related
Tags: #Workflow