📢 Marketing

Buyer Persona & Ideal Customer Profile Builder

Build research-backed ICPs and buyer personas with jobs-to-be-done, objections, buying committee, and messaging that actually converts.

0.0
0Reviews
P
August 24, 2026

Prompt

Act as a B2B customer research lead building ICPs and personas that sales, product, and marketing can use without cringing. Ground every claim in the evidence from Inputs. Label assumptions. Never invent customer quotes.

Inputs:
- Company: [What you sell, in one line]
- Offer: [Product, price band, motion]
- Best customers so far: [Firms, roles, what they bought]
- Lost deals: [Who churned or said no, and why if known]
- Sales motion: [Self-serve / sales-assist / enterprise]
- Markets: [Geo, segment]
- Evidence I have: [Calls, win-loss, usage data, or "thin"]
- Hypothesis I want pressure-tested: [Who the buyer is, or the job]

Generate:
1. Evidence grade: What is Provided vs Assumed vs Unknown. List 5 questions that would most improve this ICP. If Evidence is thin, shrink claims and expand the research plan.
2. ICP (firmographics): 2 ICP tiers (Core and Adjacent) plus an explicit Disqualify profile. For each: industry, company size, revenue or funding if relevant, tech triggers, geography, buying trigger, and why they can pay. Use ranges. Do not invent a TAM.
3. Buying committee: Table of roles: champion, economic buyer, technical evaluator, end user, blocker. For each: what they need to believe, what they fear, artifact they want (ROI one-pager, security sheet, sandbox), and veto power. Mark which roles are Assumed.
4. Deep persona (one primary): Name is a role title, not a cute fake person. Demographics only if they affect buying (seniority, scope, success metric they are on the hook for). Day-in-the-work. How they get information. Tools they already live in. Quote policy: if you need illustrative language, write it as a hypothesized thought in italics and label it "not a real quote."
5. Jobs, pains, gains:
   - Jobs-to-be-done (functional, emotional, social).
   - Pains with severity 1-5 (judgment, labeled).
   - Gains / desired outcomes.
   - Current alternatives and switching costs.
6. Objection map: 6-10 objections. For each: likely owner on the committee, underlying fear, weak reply (what average vendors say), strong reply (specific, honest, with a proof type to attach). Include price, timing, "we already have X," security, and implementation.
7. Messaging: Positioning one-liner. 3 value props mapped to jobs. Email/LinkedIn opening that names the job, not the product. Words to use / words to ban. Homepage H1 options (3). What we will not claim.
8. Interview guide to disprove the persona: 10 questions. At least 4 are falsifiers (answers that would force us to rewrite the ICP). Include a "describe the last time you bought something like this" prompt. Do not ask "would you use our product?"

Constraints:
- No fake quotes, fake names like "Marketing Mary," or fake statistics.
- Label every non-provided claim as Assumption.
- Do not flatten the committee into one persona.
- If Lost deals are provided, the Disqualify profile must reflect them.
- Tone: Research memo, not a brand workshop poster.

Instructions

Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.

Examples

Example Input

Company: B2B product analytics for product-led SaaS
Offer: $1.2k-4k/mo platform, 14-day trial
Best customers so far: Series A-C product-led SaaS, 20-80 person product+eng orgs, already on Mixpanel or Amplitude and unhappy with time-to-insight
Lost deals: enterprise IT-heavy buyers who wanted SSO + DPAs in week 1; tiny indie hackers
Sales motion: product-led with a 2-call close for teams over 15 seats
Markets: US and UK
Evidence I have: 18 closed-won notes, 7 churn interviews, no third-party TAM study
Hypothesis I want pressure-tested: the champion is a PM, the blocker is the data/analytics lead

Reviews (0)

Please login to leave a review.
Loading reviews...