How to Use STP Marketing When AI Can Write for Everyone
Segmentation, targeting, and positioning matter more when your AI agent can create endless pages. Use STP to force one segment, one target, one promise, and one measurable activation loop.
AI made generic marketing cheap.
That sounds useful until every buyer gets the same polished page with different nouns swapped in.
Claude Code, Cursor, Codex, Hermes, and similar agents can draft landing pages, docs, launch posts, pricing copy, onboarding screens, and outreach angles faster than most teams can review them.
The risk is not bad writing. Bad writing is easy to spot.
The risk is plausible writing for nobody in particular.
STP fixes that by making your agent choose before it writes:
segment truth → target choice → positioning promise → measured activationIn plain English: who is this for, why them, what should they remember, and how will we know it worked?
TL;DR: copy this to your AI agent now
Use this before your agent writes a landing page, docs page, launch post, pricing section, or onboarding test:
Use STP marketing for this product. Identify 5 real customer segments from the product, docs, landing page, competitors, and current analytics context. For each segment, explain the job, urgency, buying trigger, objection, best surface to test, and one activation signal that would prove quality. Then choose one target for this week and write one positioning promise for that target only.Then ask for the test:
For the chosen target, write one headline, one CTA, one objection-handling section, and one activation event we should measure. Keep it specific to this segment. Do not broaden the copy to include other audiences.What is in it for me?
You stop reviewing endless “pretty good” variants.
That is the win.
Instead of asking your agent for ten versions of the same vague page, you get a smaller decision:
| What improves | Why it matters |
|---|---|
| The page | It speaks to one real buyer, not every possible visitor. |
| The prompt | Your agent has a target, objection, promise, and activation signal before writing. |
| The readout | You judge whether the right people moved toward value, not whether traffic went up. |
| The next week | You know whether to keep the target, narrow it, or move on. |
That is more useful than another headline brainstorm.
STP without the workshop smell
STP means segmentation, targeting, and positioning.
For AI builders, this should not become a brand strategy ceremony. Use it as a guardrail for agent-generated marketing.
| Step | What your agent should decide | What you measure |
|---|---|---|
| Segmentation | Which real customer groups behave differently? | Source, surface, intent, objections, setup constraints |
| Targeting | Which one group are we testing this week? | Qualified activation, not just clicks |
| Positioning | What promise should this group remember? | Whether the promise moves them toward value |
No segment truth, no page. No target choice, no campaign. No activation signal, no winner.
Segmentation is not fake personas
The lazy prompt is:
Create three customer personas for our product.
You usually get names, job titles, and fake psychographics. It feels organized. It is mostly decoration.
Useful segmentation starts with differences that change behavior:
- what job the customer is trying to finish
- how urgent the problem is
- what budget or authority they have
- what setup constraint blocks them
- what objection stops them
- what first value moment proves they were the right user
Take a B2B scheduling product.
The obvious segments might be recruiting teams, sales teams, consultants, and healthcare operators. The list alone does not help. Your agent should explain what changes between them.
Bad:
Recruiter. Likes automation.
Better:
Recruiting team at a 50-person startup. They schedule many first-round interviews, lose candidates when coordination takes days, and care about fewer back-and-forth emails before the first meeting is booked.
Now the page has something to do. The headline, CTA, setup path, and activation event all change.
Targeting means one audience wins this week
AI makes it easy to keep every audience in the copy.
That is how you get this sentence:
The all-in-one platform for founders, developers, marketers, creators, teams, and operators.
Nobody chose.
A target choice should fit in one row:
| Target | Why now | Surface | Activation signal | Kill condition |
|---|---|---|---|---|
| Recruiting teams at 50-person startups | Hiring is moving fast and scheduling delays lose candidates | Landing page + demo booking flow | First interview booked without manual email coordination | Recruiters click but do not connect a calendar or book an interview |
One target. One surface. One activation signal.
If two segments need different promises, test them separately. Do not average them into one bland page.
Positioning turns the target into a promise
Positioning is what the chosen segment should remember.
Weak AI positioning often becomes a feature pile:
- automation
- scheduling
- reminders
- calendar sync
- workflows
- integrations
All true. Still not a position.
For the scheduling product, if the target is recruiting teams at 50-person startups, the promise is not:
Calendar automation for teams.
It is closer to:
Book qualified first-round interviews before candidates go cold.
Or, if the page needs to be more concrete:
Turn candidate availability, recruiter calendars, and interviewer slots into one confirmed meeting without another coordination thread.
The exact sentence can change. The discipline should not.
Use the prompt, then make the test small
The TL;DR prompt gives your agent the target and positioning promise.
Do not let it turn that into a giant campaign plan. Ask for one surface. One headline. One CTA. One objection-handling section. One activation event.
Small tests are easier to read.
Measure target quality, not copy output
STP only matters if it changes what you measure.
Do not ask whether the agent wrote a better page. Ask whether the chosen segment moved closer to value.
Good STP readouts include:
- where the segment came from
- which surface they entered through
- which CTA they clicked
- which onboarding step they reached
- whether they hit the activation event
- whether they came back after first value
- whether they upgraded or showed buying intent
If you use Agent Analytics, connect this to the closed-loop growth analysis guide and ask:
Use Agent Analytics to read the STP test for <project>. Did the chosen segment reach the activation signal we defined? Compare source, surface, CTA, and activation quality. Tell me whether to keep this target, narrow the positioning, or test a different segment next week.Common mistakes
- Letting the agent invent fake personas instead of finding real customer differences.
- Keeping every audience because AI makes variants cheap.
- Changing copy before defining activation.
- Judging by traffic instead of qualified progress.
- Keeping weak positioning because it mentions every feature.
How STP fits the series
Use Bullseye when your agent needs to choose channels.
Use AARRR when your agent needs to diagnose the growth loop after users arrive.
Use STP when your agent needs to decide who a surface is for before it writes more copy.
Read next:
- The Bullseye Method for Technical Indie Hackers for channel selection.
- AARRR for Solo Builders for weekly growth diagnosis.
- Your Product Isn’t One Website Anymore for multi-surface product context.
For setup, use the Agent Analytics Skill guide. For the product-system model behind multi-surface STP tests, read Projects, Surfaces, and Portfolios.
Final framing
AI made it easy to generate marketing for every audience.
STP is the pause before that happens.
Pick the segment. Pick the promise. Measure whether the right people moved.
Otherwise you are just producing more copy and hoping volume turns into growth.
Start free with Agent Analytics.
