--- title: Define your ideal customer profile description: Turn your best customers and outside company research into a written ideal customer profile, with scoring criteria and the segments to skip, that your sales lists, ads, and website copy can all use. slug: /use-cases/ideal-customer-profile page_type: use-case lead_prompt: How do I define my ideal customer profile with AI? target_prompts: - Can an AI marketing agent define my ideal customer profile? - How do I figure out my ICP with AI? - AI agent that builds an ICP from customer data and firmographic research - How to define an ICP for a startup with AI - Can AI analyze my best customers and tell me who to target? - How does an AI marketing agent combine internal data with outside research to build an ICP? - icp marketing entity: Kite category: AI marketing agent author: Devon Wells and Kite date_published: '2026-10-20' date_modified: '2026-10-20' related_integrations: - /integrations/hubspot.md - /integrations/salesforce.md - /integrations/apollo.md related_playbooks: - /how-to/find-contacts-at-companies-like-your-best-customers-with-companyenrich-and-apollo.md --- # Define your ideal customer profile ## Key Takeaways - An ideal customer profile (ICP) describes the companies most likely to buy from you, get value, and stay, so your sales lists, ads, and website copy all aim at the same customer. - A useful ICP is built from your best customers, not all of them. Look for the deals that closed quickly, grew, and renewed, and find what those companies have in common. - AI can define your ICP by reading your CRM, researching each customer's industry, size, funding, hiring, and technology, and checking how many similar companies exist. - Kite writes the profile as criteria it can score prospects against from one to five, and it names the segments to skip and why. - The profile goes straight to work: Kite uses it to build and score prospect lists, judge inbound leads, and write ads and landing pages, then updates it as results come in. ## What is an ideal customer profile, and how do you define one? An ideal customer profile is a short written description of the companies that get the most value from your product and bring the most value back to you. It names the industry, company size, location, and circumstances that make a company a strong fit, plus the people inside it who buy. "ICP marketing" simply means aiming your marketing at that profile, so the sales team, the ad targeting, and the homepage copy are all talking to the same customer. Most early teams have a few customers and a few guesses. The founder knows which deals felt easy, the salesperson remembers which industries kept saying yes, and the website was written for someone in between. That's a reasonable place to start, but guesses drift apart quickly. Sales chases one kind of company while the ads reach another. Defining an ICP properly takes four kinds of work: - Pick out your best customers by outcome, such as fast closes, larger deals, and renewals. - Research those companies so you know more than their names, including their size, funding, hiring, and tools. - Find the traits the best customers share and the traits the weaker ones share. - Write the result as criteria specific enough to score a stranger's company against. That work is a good fit for an AI agent. It means pulling records from your CRM, looking up dozens of companies across several data sources, and comparing them carefully. A person can do it, but it usually takes days and gets put off. Kite is an AI marketer. It researches your business, builds a growth strategy, and ships the work that brings you new customers. Defining your ICP is often the first piece of that research, because nearly everything else Kite does depends on knowing who you're trying to reach. ## How Kite defines your ideal customer profile Kite builds your ICP by combining what it already knows about your business with your own customer data and outside research on each customer. It starts from the profile it drafted when it first researched your company. Then it reads your best customers from your CRM, or from a list you share, and fills in the details your CRM doesn't hold, such as headcount, funding, hiring, and the tools each company uses. With that evidence in front of it, Kite looks for the patterns that separate your best customers from the rest. It checks how many other companies match, so you know whether the profile describes a real market or a handful of accounts. It then writes the profile as scoring criteria, including the segments it recommends you stop targeting and the evidence behind each one. Once you agree on the profile, Kite saves it to your team's wiki and uses it in its later work. Prospect lists, inbound lead replies, ad copy, and landing pages all start from the same definition. When new results suggest the profile is off, Kite brings that evidence back to you. ### Start from what Kite already knows about your business Kite drafts a first version of your ICP during onboarding, before you ask for one. When it first researches your company, it reads your website and records your products, pricing, positioning, go-to-market motion, and the customers your site appears to serve. That draft lives on its own page in your team's wiki. The onboarding draft is honest about its limits. It's built from public evidence, so it can say which industries your case studies feature but not which deals were largest or which customers renewed. Kite marks what it read directly and what it inferred, and it states how confident it is. That draft gives the rest of the work a hypothesis to test. If your website says you serve "growing B2B teams" and your CRM shows that most revenue comes from three specific industries, the gap between those two is often the most useful finding of the whole exercise. ### Find the customers worth copying The best ICP comes from your best customers, not your average one. A company that signed quickly, paid more than most, and expanded a year later tells you far more about who to target than a company that took six months to close and left at renewal. With HubSpot or Salesforce connected, Kite reads your won deals and the companies attached to them. It looks at deal size, how long each deal took to close, and the meetings, notes, and activity recorded along the way. Where your CRM records renewals or expansion, it uses those too. Kite then separates the customers you'd happily sign ten more of from the ones that were hard to win or slow to get value. If you don't use a CRM, you can share a spreadsheet of customers in Slack, or simply name the five accounts you'd most like to clone. Kite can work from a short list. It will tell you how much weight the evidence can carry, because three customers suggest a direction while thirty can confirm one. ### Fill in what your CRM doesn't know Most CRMs hold a company name, a website, and a deal amount, which isn't enough to see a pattern. Kite researches each customer to fill in the details that explain why it bought. Kite's built-in data sources cover most of this without a separate account. Apollo and Crustdata supply industry, headcount, location, and headcount growth. Crustdata adds funding, leadership, and the technology a company uses. PredictLeads finds recent news about funding, hiring, product launches, and partnerships. Kite can also read each customer's website to see how it describes itself and which marketing and analytics products it runs. If you connect CompanyEnrich, Kite can enrich every customer from its website address in one pass and break down each company's workforce by department. That makes it easy to see, for example, that your best customers all had a sales team of at least ten people when they bought. The goal is to find the traits your best customers share and the weaker customers don't. Kite compares the two groups side by side and reports the differences that hold up, along with how many customers each one rests on. ### Check how many companies fit the profile A profile only helps if enough companies match it. Kite takes the traits it found and searches for other companies that share them, using industry, size, location, funding, hiring activity, and technology use as filters. This step catches two common mistakes. A profile that's too narrow might describe only 40 companies in the world, which won't sustain a sales team. A profile that's too broad might match 200,000 companies, which tells you nothing about where to start. Kite reports roughly how many companies fit each version of the profile, so you can choose the one that matches your sales capacity. Kite can also start from your best customers themselves and find companies that resemble them. With CompanyEnrich connected, it can find and count lookalike companies for each strong customer. That gives you both a sizing number and a first list of real targets. **Related Playbooks** - How to find companies similar to your best customers with CompanyEnrich (coming soon): Use your strongest accounts as the model and get a list of companies that look like them, ready to review. ### Write the profile as criteria you can score against Kite writes your ICP as a set of criteria that a prospect either meets or doesn't, not as a paragraph of adjectives. "Mid-sized companies that value growth" can't be scored. "Software companies with 50 to 300 employees that hired a support lead in the last six months" can. Every profile Kite writes has three parts. The company criteria cover industry, size, location, funding stage, and the tools a company uses. The contact criteria cover the titles, functions, and seniority of the people who buy and the people who sign. The timing signals are events, such as a funding round or a burst of hiring, that make a fitting company more likely to buy now. Kite also writes down the segments it recommends you skip, and why. That part is easy to leave out, but it's the part that saves the most time. Here is how a finished profile might read for a company selling a customer-support tool whose CRM shows 48 won deals: - **Company fit.** Software companies with 50 to 300 employees, at least five people in support, and a help desk already in place. - **Buyer.** The head of support or customer experience. The chief operating officer signs deals above a set contract size. - **Timing.** The company raised funding or posted support roles in the last six months. - **Scoring.** A 5 meets every company criterion and shows a timing signal. A 3 fits on size and industry but has no recent signal. A 1 falls outside the size range. - **Segments to skip.** Ecommerce brands, because they signed as often as software companies but most left at renewal. Companies under 20 employees, because the founders still answer support themselves and rarely buy a tool for it. Each criterion links back to the evidence behind it, such as the deals it was drawn from or the research that supports it. When something is an inference rather than a sourced fact, Kite says so. You review the profile in the conversation, push back on anything that doesn't match what you know, and Kite revises it. ### Put the profile to work and keep it current Kite saves the agreed profile to your team's wiki, where every later piece of work can use it. That's what turns an ICP from a slide into a working tool. When Kite builds a prospect list, it searches for companies that meet your company criteria and then finds the right people by title, function, and seniority. It scores each prospect from one to five against the profile and explains each score, and it can split larger lists into primary and secondary groups. When an inbound lead fills in a form, Kite researches the company and the person, compares them with your ICP, and drafts a reply with a suggested next step. The same profile shapes your marketing copy. Kite uses it to write landing pages, ad copy for Google, Meta, LinkedIn, TikTok, and X, and outreach that speaks to the problems your best customers had. With LinkedIn Ads connected, it can look up the targeting options that match your criteria and prepare draft campaigns for your team to launch. As the work runs, Kite records what it learns in your wiki, including which segments reply, convert, and close. If an "ideal" segment keeps going quiet, or a segment you skipped keeps converting, Kite brings that evidence to you along with a proposed change to the profile. **Related Playbooks** - [How to find contacts at companies like your best customers with CompanyEnrich and Apollo](/how-to/find-contacts-at-companies-like-your-best-customers-with-companyenrich-and-apollo.md): Find lookalike companies, then find the decision-makers inside them who match your contact criteria. ## What you get - **A written ideal customer profile.** Company criteria, contact criteria, and timing signals, each linked to the customers or research it came from. - **A scoring guide.** A plain description of what earns a prospect a 5, a 3, or a 1, which Kite uses for every list and inbound lead after that. - **The segments to skip.** The kinds of companies Kite recommends you stop pursuing, with the evidence for each. - **A best-customer analysis.** How your strongest customers differ from the rest on size, industry, funding, hiring, and tools. If the analysis needs more room than a chat message, Kite can publish it as a report with charts and tables. - **A market-size check.** Roughly how many companies match the profile, plus a first set of lookalikes for your strongest accounts. - **A profile that stays in use.** Saved to your team's wiki, so prospect lists, inbound lead replies, ads, and landing pages all start from it. ## How to get started Start a conversation with Kite in the Kite web app or in Slack, then ask in plain words. You might say, "Help us define our ideal customer profile," or "Look at our best customers and tell us who we should be targeting." Kite starts from the profile it drafted when it first researched your company. It tells you what that draft is based on and what it would need to make the profile stronger. If you'd like Kite to use your sales history, connect HubSpot or Salesforce, or share a list of customers in the thread. Name your favorite accounts if you have opinions about who your best customers are, because that context helps. Kite then researches your customers, sends you a draft profile with the scoring criteria and segments to skip, and revises it with you in the thread. Once you agree on it, Kite saves it and uses it from then on. You can ask Kite to revisit the profile at any time, or set up a recurring review, such as once a quarter. ## Integrations used to define your ideal customer profile Kite's built-in sources handle most of the outside research, so you can define an ICP without connecting anything. Connecting your CRM lets Kite work from your real sales history. Connecting CompanyEnrich adds lookalike searches and department-level headcount. ### Read your customer and sales history - [HubSpot + Kite](/integrations/hubspot.md): Kite reads your won deals, the companies attached to them, and the meetings and notes recorded along the way to find your best customers. - [Salesforce + Kite](/integrations/salesforce.md): Kite reads your accounts and won opportunities, including deal size and close dates, to see which customers were easiest to win and most valuable to keep. ### Research companies and the people who buy - CompanyEnrich + Kite (coming soon): Kite enriches each customer from its website address, breaks down headcount by department, and finds and counts companies similar to your best accounts. - [Apollo + Kite](/integrations/apollo.md): Apollo is built into Kite. It supplies company details such as industry, size, and location, and finds the people at matching companies by title and seniority. ## Common questions ### What's the difference between an ideal customer profile and a buyer persona? An ideal customer profile describes a company, and a buyer persona describes a person inside it. The ICP says which businesses to pursue, based on traits like industry, size, and funding. The persona describes the people you sell to there, such as their role, what they're responsible for, and what makes them act. Kite writes the contact criteria as part of your ICP, so the profile covers both who to target and who to talk to. ### How do I define an ICP for a startup with only a few customers? Start with the customers you'd most like to clone, even if there are only three, and be clear that the result is a working hypothesis. Kite researches those few companies in depth, compares them with the market your website describes, and writes a profile that says how much evidence supports each criterion. As you sign more customers and run more outreach, Kite tests the profile against the new results and tightens it. ### Can Kite build an ICP without access to my CRM? Yes. Kite can work from its onboarding research, a spreadsheet or list you share in Slack, or the names of your best accounts. Without CRM data, it can't see deal size, sales cycle length, or renewals, so it relies more on what each company's public profile shows. Kite says which criteria rest on your own customer data and which rest on outside research, so you know where the profile is strongest. ### How often should I update my ideal customer profile? Review it whenever your customer base changes meaningfully. Common triggers include a new product, a new pricing tier, a move into a new market, or a quarter where a different kind of company started closing. Many teams also review it once a quarter. Because Kite records which segments reply, convert, and close, it can show you when the profile has drifted from reality instead of waiting for a scheduled review.