Pricing decisions usually fall out of a Slack thread and a gut feeling. A memo forces the assumptions to the surface: who the buyer is, what the anchor is for, why the middle tier is where 70% of revenue should land, and how low you will go before walking away. AI is good at drafting the memo; you are good at deciding whether to ship it.
The task
You need to set or reset pricing. You have a product, a rough sense of three buyer segments, and competitor screenshots open in another tab. You want a memo your team can fight over before Friday: three tiers with named differentiators (not “more seats”), the role of each tier (anchor, hero, ceiling), a single discount floor you will not cross in sales calls, and one experiment to run in 30 days to test the call.
When this is the right job for AI
AI is strong at structuring the memo, naming the role each tier plays, and pulling pricing-psychology language (“anchor,” “decoy,” “tier compression”) into a working draft. It is weak at telling you what the market will actually pay — that is a function of your buyer interviews, win/loss data, and competitor benchmarks. Hand the model your data; let it shape the argument. Do not let it invent the willingness-to-pay numbers.
What to feed the AI
- The product and the unit of value you charge for (seats, usage, outcome, hybrid)
- Three buyer segments with one-line descriptions and their current alternative
- Competitor pricing (at least three, with their tier names and price points)
- Your current pricing if you have one, and what you know is broken
- One willingness-to-pay data point per segment (from interviews, not guesses)
- Your gross margin so the discount floor is grounded in unit economics
- The decision deadline and who has to approve
Copy-ready prompt
You are drafting a pricing strategy memo for an internal pricing review.
Product: <one line>
Unit of value we charge for: <seats / usage / outcome / hybrid>
Buyer segments (3):
Segment A: <description> | current alternative: <X> | WTP signal: <Y>
Segment B: ...
Segment C: ...
Competitor pricing (at least 3): <name | tiers | prices>
Current pricing (or "none"): <paste>
Known issues with current pricing: <list>
Gross margin: <%>
Decision deadline: <date>
Return a memo with these sections, exactly in this order:
1. The pricing question we are trying to answer (one sentence).
2. Three proposed tiers, each with:
- Tier name (no generic "Pro / Business / Enterprise" unless it fits)
- Headline price + billing unit
- Role: anchor / hero / ceiling
- Two named differentiators that a buyer can verify in 10 seconds
- The segment this tier is built for
3. Anchor logic: why the top tier price is credible and not laughed at.
4. Hero tier logic: why 60-70% of revenue should land here.
5. Discount floor: the single percentage off the hero tier we will not
cross. Justify against gross margin and segment WTP.
6. Three objections the team will raise, with the counter for each.
7. One 30-day experiment to test the riskiest assumption — sample,
measurement, decision rule.
Do not invent WTP numbers. If a segment is missing a WTP signal,
flag it as [needs interview] in the memo.
Tone: argumentative but specific. No "value-based pricing" without
naming the value.For the second pass: “Now write the one-paragraph version a CEO would forward to the board.”
Sample output structure
Question: Should we collapse from five tiers to three and raise the hero tier from $49 to $79?
Tier 1, Maker, $19/seat/mo (anchor): for solo creators, role is to make the hero tier feel obviously priced. Differentiators: 1 workspace, community support.
Tier 2, Studio, $79/seat/mo (hero): for 3-15 person teams. Differentiators: shared workspaces, audit log, priority support. This is where 65% of revenue should land.
Tier 3, Agency, $249/seat/mo (ceiling): for client-services teams. Differentiators: white-label, SSO, dedicated CSM.
Discount floor: 15% off the Studio annual list. Below that we are inside the LTV/CAC danger zone for the segment.
How to refine
- If tier differentiators are generic: “Replace ‘more storage’ and ‘priority support’ with a named capability a buyer can verify. ‘Audit log’ is verifiable; ‘better support’ is not.”
- If the anchor reads cynical: “The anchor tier needs at least one buyer who would genuinely choose it. If you cannot name one, the anchor is decoy-only and that will leak in sales calls.”
- If the discount floor is missing a justification: “Tie the floor to gross margin or segment WTP. ‘We feel uncomfortable below 20% off’ is not a floor.”
- If objections are softballs: “Surface the real objection from sales — ‘we will lose deals at $79’ — and write the counter that gives sales a script, not a brush-off.”
Common mistakes
- Letting AI invent willingness-to-pay numbers without source data
- Naming three tiers but designing only the hero. Anchor and ceiling matter equally
- A discount floor with no tie to unit economics. Sales will breach it the first quarter
- “Enterprise: call us” with no anchor price. Buyers self-select out before the call
- Forgetting the experiment. A memo without a 30-day test is a manifesto, not a decision
FAQ
- Three tiers or four?: Three unless you have data showing a real 4th segment. Four tiers usually means two are doing the same job.
- Per-seat vs usage?: Match the unit to the value the buyer feels grow. Per-seat is fine for tools used daily; usage is fine for tools whose value scales with throughput.
- When to re-price?: Annually, or after any 20% shift in CAC, churn, or competitive landscape. Re-pricing more often erodes trust.
- How do I know the hero tier is right?: 60-70% of new revenue lands there within two quarters; if it is below 40%, the differentiation is unclear.
Related
- Pricing experiment AI — design the 30-day test
- Pricing hypothesis — earlier-stage pricing reasoning
- Pricing page copy — translate the memo into landing-page copy
- Unit economics AI — the numbers your discount floor is anchored to
- Board deck narrative AI — pricing usually shows up as a slide
- AB test summary AI — read out the experiment results