How I Build ICPs with Claude (And Why Most Teams Get Them Wrong)
Most ICPs are fictional. A few adjectives, a job title, some assumed pain points. Written once in a strategy offsite. Never updated.
The result: your sales team targets the wrong accounts, your messaging stays generic, your win rate doesn’t move. Not because the product is wrong. Because the ICP was never grounded in anything real.
Here’s the methodology I use to fix that. If you want to understand why I am doing this with claude, check Claude vs. ChatGPT for GTM.
Why Most ICPs Fail Before They’re Used
A typical ICP document looks like this:
- Industry: SaaS, 100–500 employees
- Title: VP of Sales or CRO
- Pain: Struggling with pipeline visibility
- Budget: $50k–$200k
That’s a market segment, not an ICP. It tells you who could buy. It says nothing about who actually buys, stays, and expands.
The difference matters more than most teams realise. A market segment tells you where to look. An ICP tells you what you’re looking for when you get there. Without that precision, your reps are qualifying on gut feel, your messaging is broad enough to say nothing, and your win rate reflects it.
In How to Build a GTM Strategy Step by Step, I covered the four elements that make an ICP actionable. That’s the foundation this workflow builds on.
The Four Elements That Make an ICP Actionable
Before touching Claude, you need to know what you’re building toward. An ICP is only useful if it answers four specific questions.
1. The Trigger
What is happening in their world that makes them need a solution right now? Not eventually. Now.
In B2B, it’s usually an event: a new hire, a failed launch, a missed quarter, a contract renewal that forces a decision. In FMCG, it’s a consumption moment: the specific occasion or job-to-be-done that drives the purchase. For a full FMCG GTM framework built around these dynamics, see Go-to-Market Strategy for FMCG: A Practical Guide.
This is the most important element and the one most ICPs get wrong. “Struggling with pipeline visibility” is not a trigger. “New VP of Sales in seat for 60 days, under pressure to show results before Q2 board review” is a trigger. One tells you nothing about when to reach out. The other tells you exactly when the window opens.
2. The Cost of Inaction
What does it cost them if the problem stays unsolved? Time, money, competitive position, internal credibility.
Quantify it where you can. Vague pain is hard to sell against because the prospect can always decide to live with it. Specific pain has a price tag. When someone knows that the status quo costs them $200k a year in wasted sales capacity, the conversation about your $50k solution gets much easier.
If you can’t articulate the cost of inaction clearly, you don’t have a sales motion yet. You have a feature list.
3. The Decision Dynamics
Who uses the product, who pays, who can block the purchase? In B2B, these are rarely the same person. The user wants the tool. The buyer controls the budget. Legal or IT can kill the deal at the last minute.
In consumer and FMCG, it’s more varied than it looks. Gifting, foodservice, and family purchase decisions all involve stakeholders with different priorities. Mapping this correctly tells you who to reach, with what message, at what point in the buying process.
I’ve seen deals die because the sales process was built entirely around the user champion and ignored the economic buyer. The ICP needs to account for both.
4. The Alternatives
What are they doing today instead of buying from you?
This is where most ICP exercises stop too early. Teams list competitors. The real answer is usually messier: spreadsheets, manual processes, a cheaper tool that’s 70% good enough, or nothing at all because the problem hasn’t reached the threshold of pain that justifies a purchase.
Understanding the status quo tells you what you’re actually competing against. It also reveals the real objection you need to overcome, which is almost never “your product isn’t good enough.” It’s “the cost and friction of switching isn’t worth it yet.”
Why the Order Matters
The four elements aren’t a checklist you can fill in any order. They build on each other.
Trigger comes first because it defines the buying window. Without a clear trigger, you don’t know when to reach your ICP. Your outreach becomes constant noise instead of well-timed signal.
Cost of inaction comes second because it defines the urgency inside that window. A trigger opens the door. The cost of inaction determines how wide it opens and how long it stays open.
Decision dynamics come third because they tell you who to engage once you’re in the window. Knowing the trigger and the pain is useless if you’re talking to the wrong person.
Alternatives come last because they define the real competitive context. This is where your positioning gets sharp or stays generic.
Teams that skip directly to firmographics are answering “who” before they’ve answered “when”, “why”, and “how”. The result is an ICP that describes a lot of companies but qualifies none of them well.
Before You Build: Data Quality
Claude can only synthesize what you give it. The quality of your ICP is a direct function of the quality of your input data.
The minimum you need:
- 10–20 closed-won deals from the last 12 months
- 5–10 churned or lost deals
- CRM fields: industry, company size, deal cycle length, ACV
- Any customer interview notes or call recordings (summarised is fine)
If your CRM data is messy (missing fields, inconsistent industry tagging, no close reasons logged), the output will reflect that. That’s not a reason to wait. It’s a reason to run the analysis now, flag the gaps explicitly, and use the output to tell your RevOps team exactly what to start capturing.
An ICP built on imperfect data that acknowledges its limitations is more useful than no ICP. The ⚠️ flag system in the prompt below forces those limitations into the open rather than letting Claude paper over them with confident-sounding generalisations.
The Core Prompt: ICP Draft
Once you have pattern data from your deal history or customer interviews, this prompt builds the ICP. It structures the output directly around the four elements above and forces specificity at every step.
You are a B2B GTM strategist.
Using the pattern analysis above, draft a complete Ideal Customer Profile. Structure it around these four elements:
1. Trigger — what specific event or situation makes them need a solution right now?
2. Cost of inaction — what does it cost them (time, money, reputation) if the problem stays unsolved? Quantify where possible.
3. Decision dynamics — who uses the product, who pays, who can block the purchase?
4. Alternatives — what are they doing instead of buying from us today?
Also include:
– Firmographics: industry, company size, revenue range, geography
– Champion profile: role, seniority, department of the internal driver
– Exclusion criteria: who should we explicitly not pursue, and why?
Rules:
– No vague language (“growing company”, “innovative team”). Name actual signals.
– Every criterion must be assessable from public data or a 15-minute discovery call.
– Flag any criterion where the source data was thin or absent (mark with ⚠️).
Two things to watch for in the output. First, if Claude returns vague answers on trigger or alternatives, the source data is thin in those areas. That’s a signal to dig deeper before building anything around the ICP. Second, pay attention to the ⚠️ flags. They’re not a sign of failure. They’re a list of the gaps your next round of customer conversations should close.
When Is an ICP Finished?
An ICP is never finished. But it’s ready to use when it can answer three practical questions without hedging.
Can a rep look at a company’s LinkedIn page and publicly available data and assign a preliminary qualification score in under five minutes? If the answer is no, the criteria are too vague or too dependent on information you can only get in a discovery call.
Can a new hire read it in 60 seconds and know exactly who to call and who to skip? If it takes longer than that, or if they’d still be unsure, the document is too long or too abstract.
Would two different reps looking at the same account reach the same qualification decision? If not, the criteria aren’t specific enough to drive consistent behaviour.
When the ICP can pass all three, it’s ready. Until then, it’s still a hypothesis.
When to Refresh
Most teams treat ICP as a one-time exercise. That’s why their targeting drifts and their win rates plateau.
Refresh quarterly, and also whenever:
- Win rate drops more than 10 percentage points
- You enter a new segment or launch a new product line
- Churn patterns shift noticeably
- Deal cycles start getting consistently longer
The refresh isn’t a full rebuild. It’s a delta analysis: which criteria held up against new data, which weakened, what new patterns emerged. If you documented the previous version well, the diff between versions is often the most useful output. It shows you how your ideal customer is evolving and where your GTM assumptions need to catch up.
FAQ
Ten closed-won deals from the past 12 months is enough to find initial patterns. More data improves signal quality, but don’t let the perfect dataset stop you from starting.
Yes, but be explicit about the limitation. Flag which criteria are hypotheses rather than evidence. Treat the output as a testable assumption, not a finished ICP. The value at early stage is in making the assumptions visible so you can test them deliberately.
A buyer persona describes who a person is: demographics, motivations, communication style. An ICP describes which companies to sell to and under what conditions. Both are useful. The ICP comes first because it determines who you’re talking to before you decide how to talk to them.
The free tier handles shorter prompts. For multi-step workflows with larger datasets, Claude Pro manages longer context more reliably.
The Prompts
The methodology above tells you what to build and why and the prompt handles the ICP draft itself.
I discuss the full ICP workflow (pattern extraction from CRM data or customer interviews, segment comparison, pressure-testing, quarterly refresh, scoring model, discovery call guide, and pre-call account brief) in the ICP Prompt System. Four stages, twelve prompts, one per stage shown free, all twelve in the pack that can be downloaded for free.
Related:
What is a Go-to-Market Strategy? (ICP in the context of a full GTM system)
How to Build a GTM Strategy Step by Step (the four-element ICP framework this article is built on)
The ICP Prompt System (the 12 prompts that operationalise this methodology)
