persona-architect: 5 deep buyer personas
What it does
Section titled “What it does”persona-architect builds 5 deep buyer personas for your market. Not demographics, not age ranges and income brackets. The real stuff: what each persona is afraid of, what triggers them to buy, what objection kills the sale, and the exact language they use about their problem.
Each persona covers psychological triggers, a cognitive bias profile, decision-making style, layered pain points (surface, underlying, emotional root), and objections with counters. The output feeds directly into copywriting, ad creative, and page builds, so every downstream asset is aimed at a real human, not a composite spreadsheet.
It auto-runs apollo first to gather the customer intelligence it needs. If you’ve already run apollo in the same session, it picks up from that research instead of starting over.
When to reach for it
Section titled “When to reach for it”- You’re starting work for a new brand and need to understand who actually buys.
- You’re about to write copy or build a page and want it grounded in real psychology, not assumptions.
- Your ads work for one segment but fall flat everywhere else and you want to find the other buyers.
- You’re briefing a creative team and need clear, distinct archetypes to write to.
- You’ve run apollo and want to turn the intelligence into actionable personas.
Where it runs
Section titled “Where it runs”persona-architect works in both environments. The research phase needs a live web connection (both Cowork and the desktop app have one). The persona development is pure reasoning, so it runs cleanly either way.
How to run it
Section titled “How to run it”Point it at a brand:
“Run persona-architect for examplebrand.com”
Or pair it with an apollo run:
“/um-toolkit:apollo full intelligence on examplebrand.com, then build personas with persona-architect”
If research is already in context, it skips apollo and goes straight to persona development:
“We just finished the apollo research. Now run persona-architect on that.”
After building all 5, it presents them and asks which to keep. Take individual ones (“1, 3”), all of them, or none if you want to steer differently.
What it needs from you
Section titled “What it needs from you”- A brand URL or a clearly described market. The more specific, the sharper the personas.
- (Optional) Any existing customer research: survey results, reviews, or sales call notes. Hand it over and the personas ground in your actual data, not just what’s publicly findable.
What you get back
Section titled “What you get back”5 distinct persona profiles, each covering:
- Core identity: name, archetype, age range, occupation, and a one-sentence summary of who this person is.
- Awareness level: where they sit on the Schwartz spectrum (unaware, problem-aware, solution-aware, product-aware, most aware), which changes how you talk to them.
- Layered pain points: the surface complaint they’d tell a friend, the underlying frustration they’d admit privately, and the emotional root they might not name to themselves.
- Decision triggers: the specific conditions that make this persona ready to buy.
- Cognitive bias profile: the top psychological levers that move them, with notes on applying each in creative.
- Primary objection and counter: the one thing most likely to stop the sale, and the reframe for it.
- Voice-of-customer language: exact phrases from research that sound like this persona talking.
At the end, a quick reference matrix puts all 5 side by side so you see the differentiation at a glance.
A worked example
Section titled “A worked example”You: “Run persona-architect for examplebrand.com (sleep supplements).”
persona-architect: (after apollo research)
Persona 2 of 5: The Wired Professional
Age 38, operations manager, two kids, perpetually behind on sleep. Surface pain: “I can’t wind down.” Underlying pain: melatonin stopped working after four tries, so she doesn’t trust anything. Emotional root: she’s embarrassed she can’t solve something so basic.
Decision trigger: a specific mechanism she hasn’t tried, explained without hype.
Primary objection: “This is just melatonin dressed up.” Counter: lead with the ingredient differentiation, prove the mechanism before promising outcomes.
Cognitive bias: authority and specificity effect. She responds to clinical explanations with real dosage numbers, not lifestyle photography.
Voice of customer: “I’ve tried everything. I just want something that works without making me groggy in the morning.”
Each of the 5 personas is this specific, and each differs meaningfully in awareness level, decision style, and the creative approach that reaches them.
Tips & gotchas
Section titled “Tips & gotchas”- The 5 personas are designed to be different. Each gets a different awareness level, primary pain, and decision style. If they start to blur together, the research wasn’t deep enough. Run apollo first with full intelligence.
- Let it run apollo if you haven’t already. Skipping research produces personas that sound plausible but aren’t grounded. The output is only as good as the input.
- Persona count is flexible. The default is 5. Tell it you want 3 for a simple market, or up to 8 for a complex multi-segment one.
- The quick reference matrix is the working doc. The side-by-side matrix is what you’ll actually refer back to when briefing creative or choosing a persona to write an ad for.
- It won’t invent data. If research didn’t surface enough language for a persona, it flags the gap with a placeholder rather than filling it with something plausible-sounding.
Related skills
Section titled “Related skills”persona-architect is the bridge between market research and creative execution. Run apollo first for the raw intelligence, then persona-architect to turn it into people you can write to. From there, hermes uses the profiles to generate copy and meta-image-ad-generator uses them to build ad creative aimed at specific archetypes.