Strategy & ICP

Discover your ideal customer

Work out who could benefit, what you’re assuming, and what to test next—with your agent.

Copies a prompt for your AI assistant

View prompt

Start wherever you are

Bring a rough idea or an existing profile. Your agent helps you take the next step.

Start with your AI assistant. Connect tools when needed.

Fictional example

You

I want to help teams stop losing client details.

Agent

Where have you noticed that happening, and who was involved?

You

I noticed it in my design agency, when sales handed a project to delivery.

One possible direction after exploring the idea together

Draft hypothesis · ready to test

A clearer place to start.

Possible customers: Design agencies handing client work from sales to delivery.
Problem to explore: Client commitments get lost between the people making the sale and the people doing the work.

What we know

In this example, the customer has experienced this in their own agency.

Still an assumption

Other agencies experience it often enough to want help.

A next step to learn

Ask another agency to walk through a recent client handover.

Context & open questions

Look for: Regular handovers between sales and a delivery owner.
Exclude for now: Solo operators with no handover between people.
Still open: How often context is lost, its impact, current workarounds, buying roles, and willingness to pay.

Your result will reflect your conversation. Keep it in your assistant or a document you choose.

Read the full guide

Start wherever you are

A rough idea is enough. You don’t need a customer list, a target audience, or customer evidence to begin. Your agent will help you clarify the idea through conversation. If you already have an ICP, customer examples, notes, or research, you can use them. They’re optional, and any assumptions will stay clearly labeled.

Figure it out together

Start with what you know
Your agent builds on your idea, profile, or brief and asks only what helps the next step.
Explore possible customers
Consider options together and compare their problems, urgency, and buying context.
Shape a working profile
Describe who could benefit. Separate assumptions from evidence and leave unknowns open.
Choose what to test
Agree a profile and simple customer conversations or observations to test it. Keep a copy, or optionally save it with a suitable connection.

Tools and connections

Your AI assistant: The conversation and the customer profile.
Documents or CRM · optional: Customer, account, and sales evidence from your existing tools, including Capable.

What you’ll keep

A useful result and a copyable working brief with your decisions, reasons, evidence, assumptions, actual status, pending actions, and next question. Bring it back when you’re ready to continue.

A useful first outcome

You have a plausible customer hypothesis, a clear view of what’s unknown, and a simple way to test it. The profile is ready to learn from, rather than proven.

When to revisit it

Return after conversations with possible customers, useful observations, early sales results, or changes to your idea. Use what you learn to adjust the profile and choose what to test next.

View the setup prompt

Copy into your AI assistant to begin

Help me discover or refine my ideal customer through a guided conversation. Work it out with me rather than asking me to arrive with the answer.

Begin by checking the context I have already supplied in this conversation. If I have not described an idea, product, service, problem, or existing profile, ask: "What are you working on, or what problem would you like to help solve?" If I have supplied a rough idea, acknowledge it briefly and ask the next useful unanswered question. Do not ask me to repeat what you already know.

If I bring an existing profile, plan, or working brief, build on it. Check what has changed where relevant, then continue with the next useful question. If my focus is unclear, you might ask: "What feels least certain about this profile?" Treat that as an example, not a mandatory opening. If I have already said what I want to improve, start there.

A rough idea is enough. Do not require an existing customer list, a defined target audience, a finished offer, or customer research to start.

Ask one clear question at a time and adapt to my answers. Ask only what is needed for the next useful step. Reuse context I have already shared and check what has changed. If I already have a profile or plan, help me review and refine it without restarting discovery. If I am unsure, help me think it through with concrete examples or a few possibilities to react to. Label your suggestions as possibilities, not facts about my business.

Build on the context I have shared and explore only the gaps relevant to our next step. Where useful, clarify the value the idea could provide, who might experience the problem, in what situation, how they deal with it today, and what could make solving it important enough to act on. Do not turn these topics into a required questionnaire.

Do not start by asking me who my ideal customer is. When we are choosing an audience, help me compare plausible customer groups before narrowing the focus. When we are refining an existing profile, focus on the uncertainties we choose to explore. Explain the reasoning, invite me to disagree, and summarize what we are learning along the way. Do not treat my preference for an option as evidence that it is a good market.

If I already have customers, an ICP, interviews, or sales evidence, offer to use them to refine the discussion. These are optional. Ask which information I want to use before drawing on connected documents or CRM records, including Capable. Do not make connecting tools a prerequisite.

Explore buying roles and purchase context where relevant. Do not assume an industry, company size, geography, or business model. Separate essential fit criteria from preferences and exclusions, and leave unsupported criteria open. Do not invent customer facts, arbitrary thresholds, or numerical fit scores.

Offer a provisional profile early, as soon as we have enough for something useful to react to. Do not wait for every question to be answered. Summarize a working ICP containing:
1. Who could benefit and the situation they are in.
2. The problem or outcome that could make the offer valuable.
3. Initial fit criteria, preferences, and exclusions where justified.
4. Buying roles and context, where relevant.
5. The evidence behind our choices, clearly separated from assumptions.
6. Open questions, plausible alternatives, and what we should learn next.

When there is little evidence, call this an initial hypothesis. A conversation can help us choose a direction to investigate; it does not prove demand. If we cannot narrow the audience yet, preserve the candidate groups and explain what would help distinguish them.

Review the draft with me. Agree on a small next step to test the most important assumptions, such as a customer conversation or checking how people currently solve the problem. Do not conduct research or contact people automatically.

At a useful stopping point, provide a concise, copyable working brief containing the profile, decisions and reasons, evidence separated from assumptions, unresolved questions or alternatives, actual completion status, and the next useful step and open question. Include an owner or review point only where agreed. Make clear whether research, saving, or other actions actually happened or remain pending. If I pause before we have a profile, preserve what we have learned without filling the gaps with invented answers.

Review the brief with me and ask where I want to keep it. Save it there only if I choose that and connected tools support the action; otherwise leave the copyable version in this conversation. Explain what new evidence would make it useful to revisit. The brief should let me resume here or bring selected context to another assistant. Include a continuation instruction to check what has changed, reuse the context, and start with the next useful question. Do not imply automatic memory across conversations or treat the brief as authorization to execute actions.

Keep this task focused on discovering and defining the ICP. Do not source prospects, enroll contacts, or start outreach.

How did it go? Optional feedback

Discover your ideal customer

Start wherever you are

A clearer place to start.

What we know

Still an assumption

A next step to learn

Example working brief

Start wherever you are

Figure it out together

Tools and connections

What you’ll keep

A useful first outcome

When to revisit it