Ideal Customer Profile

Define the customer most likely to find value, retain, and expand — but build it from real won-and-happy customers in the corpus, not an idealized composite. The Evermuse twist: you identify the accounts that actually bought and stayed, read their conversations for why, and let the patterns across them define the ICP with quotes attached.

Step 0 — Relevance & availability

Confirm this is an ICP / targeting task and Evermuse is connected. If the tools aren't present, produce a framework-only ICP labeled ⚠ ungrounded — Evermuse not connected and tell the user to authorize the MCP. See ${CLAUDE_PLUGIN_ROOT}/skills/using-evermuse/SKILL.md Step 0.

Evermuse Grounding (required)

Follow ${CLAUDE_PLUGIN_ROOT}/skills/using-evermuse/SKILL.md.

Required — one parallel batch of searches. Verify the product (Rule 1), then fire in a single parallel batch:

• 3–4 evidence searches, each worded from a different angle (limit up to 50). Evidence comes back rich and varied — expect large, useful result sets.
• one guidance search and one context search (limit up to 50). These are usually sparse or empty; run them anyway and note when they're thin.

Read each response's digest — it reports how many more results exist. Use judgment on whether a query is worth pulling deeper (raise limit toward the 100 max and/or page with next_offset to avoid repeats), weighing payload size, remaining context, task complexity, and the value of the data. For deep pulls, consider spawning sub-agents — instruct them to return every citation with the same metadata the tools return (url, who_said_it, meeting_name, created_at) so you can still cite.

Optional — considered use. Once grounded, reach for the other tools only when they add value: a quote-angled search for verbatim voice, read_source for a single deep dive, inline citations (references/citations.md), and add_source to save the deliverable. Sources, citations, and saving are optional — not required.

Instructions

You are defining the ICP for $ARGUMENTS. Where the source method reaches for "PMF survey data / win-loss / interviews," use the Evermuse corpus of real customer conversations as the primary source — start from accounts that actually won and stayed, and read their words.

Step 1 — Find the real best customers

Segment the accounts in the corpus by value signal: bought and renewed, expanded, fastest time-to-value, most enthusiastic, best reference potential. These accounts — not a hypothetical persona — are the raw material.

Step 2 — Extract the four ICP components (each cited)

For the best-customer cohort, synthesize patterns and attach quotes:

• Firmographics — company size, industry/vertical, geography, buyer's role/department, company stage. Note the actual common denominators, including non-obvious outliers that are high-value.
• Behaviors — how they discovered the product, evaluation process and timeline, who was in the buying committee, adoption speed and breadth, support needs. (Discovery channel is gold for GTM — cite it.)
• Jobs to Be Done — the functional job, plus emotional ("how they want to feel") and social dimensions, and their own success metrics — in their words.
• Needs & pains — the before-state frustrations and the after-state they got, quantified where the evidence allows (cost/time of the problem).

Step 3 — Sharpen the edges

• Ideal-of-the-ideal: the tightest sub-pattern that predicts the best outcomes.
• Disqualification criteria: who is NOT a fit — grounded in accounts that looked similar but churned, stalled, or never activated. This is where evidence prevents an over-broad ICP.
• GTM implications: what the ICP means for messaging and channels (hand off to /evermuse:positioning-and-messaging or /evermuse:gtm-strategy).

Deliverable format

# Ideal Customer Profile — [product]

**Built from:** [N] won-and-happy accounts in the corpus ([examples]).

**Firmographics:** [size / industry / geo / role / stage] — pattern evidenced by [`1`](URL).
**Behaviors:** discovered via [channel] [`2`](URL); buying committee [who]; adoption [speed].
**Jobs to Be Done:** functional — [job]; emotional — [feeling]; social — [status]. > "[JTBD quote]" — [attribution] [`3`](URL)
**Top pains → outcomes:**
| Before (pain) | After (outcome they got) | Evidence |
|---|---|---|
| [pain] | [outcome] | > "[quote]" — [attribution] [`4`](URL) |

**Ideal-of-the-ideal:** [tightest high-value sub-pattern].
**Disqualification (NOT a fit):** [traits] — grounded in [churned/stalled accounts] [`5`](URL).
**GTM implications:** [messaging + channel notes].

Honesty when evidence is thin

If only a handful of happy accounts exist, say so — the ICP is a hypothesis from N accounts, not a law. Flag where more won-customer conversations would sharpen it.

Save & hand off

After saving, offer: "Position for this ICP?" (/evermuse:positioning-and-messaging), "Plan the launch?" (/evermuse:gtm-strategy), or "Narrow to a first beachhead?" (/evermuse:beachhead-segment).

Further reading

• Adapted from phuryn/pm-skills ideal-customer-profile (MIT); grounding is Evermuse-native.
• Based on Jobs to Be Done theory (Clayton Christensen).
• How to Design a Value Proposition Customers Can't Resist?