Build buyer personas for messaging and campaigns

Build buyer personas from what your real customers said on calls, in chat, in your CRM, and in public reviews, so every claim in a persona traces back to a source and your team can write messaging and campaigns from it.

Key Takeaways

  1. Your target customers are the specific people most likely to buy from you and get value from what you sell. A buyer persona is a written description of one kind of target customer: their role in the purchase, what made them start looking, what it costs them, what almost stopped them, and the words they use.
  2. To build buyer personas with AI, give the AI real customer evidence to read, such as sales call transcripts, support conversations, CRM records, and public reviews. An AI asked to imagine your customer will invent one, while an AI that reads your customers can quote them.
  3. Kite reads your Fathom calls, Intercom conversations, and HubSpot records, plus public reviews and discussions where your buyers write, and pulls out each buyer's trigger, cost, objections, and exact phrases with the source of every quote.
  4. Kite groups customers by why they buy rather than by age or job title, labels every line of a persona as a quote or an inference, and lists what the evidence can't yet support.
  5. Kite saves the personas where its own writers read them before writing an ad, a landing page, or an email, and keeps them current as new calls and conversations come in.
On this page

Who are your target customers, and what makes a buyer persona useful?

Your target customers are the people most likely to buy what you sell, get real value from it, and stay. For a business that sells to other companies, that usually means two layers: the kinds of companies that fit, which is your ideal customer profile, and the people inside those companies who find, choose, approve, and use your product. For a business that sells to consumers, it's the groups of people who share the same problem and the same reason to act. A buyer persona describes one of those groups closely enough that anyone writing for your company knows what to say to them.

A persona earns its place when it changes what your team writes. A useful buyer persona covers:

  • who this person is in the purchase, and whether they pay for it, use it, or approve it
  • the event that made them start looking, such as a new hire, a lost client, or a bad month
  • what the problem costs them in hours, money, or standing with their boss
  • what they tried first, and why it fell short
  • the objections that almost stopped them from buying
  • the outcome they hoped for, described in their own words
  • the exact phrases they use for the problem, which are the lines your headlines and ads should borrow

Most personas fail because they're written from guesses. A team gathers in a room, names the persona "Marketing Mary," gives her an age, two kids, and a favorite podcast, and never opens the document again. AI persona generators make the same mistake faster. They produce a confident, detailed profile in seconds, but nothing in it came from your customers, so nothing in it tells you what your customers will respond to.

Personas built from evidence work differently. You read what real customers said on sales calls, in support chats, in reviews, and in the notes your team keeps, and you look for the patterns that repeat. That reading is slow for a person, because a single month of sales calls can run to dozens of hours. It's a good fit for an AI agent, which can read every transcript and conversation, keep each quote tied to its source, and come back with the patterns.

Kite is an AI marketer. It researches your business, builds a growth strategy, and ships the work that brings you new customers. Here's how it builds buyer personas you can use.

How Kite builds buyer personas from real customer evidence

Kite starts with what it already knows about your business. When you sign up, it researches your company and records your positioning, products, and the customers your website appears to serve in your team's wiki. That gives it a first idea of who your target customers are, which the persona work then tests against your actual customers.

Next, Kite gathers what those customers said and did. It reads your recorded sales and customer calls in Fathom, your chat and support conversations in Intercom, and the deals, notes, and contacts in HubSpot. It adds public reviews and discussions from the sites and communities where your buyers write, and it reads the LinkedIn profiles of the people who bought to see their roles and what they post about.

From that evidence, Kite pulls out the moments that matter, groups customers by why they buy, and writes a persona for each group. Every quote carries its source, every inference is marked as one, and each persona ends with what Kite still doesn't know. Once you agree on the personas, Kite saves them to your wiki, uses them in the ads, pages, and emails it writes, and updates them as new evidence comes in.

  1. Start from the customers who bought
  2. Gather what customers actually said
  3. Pull out the moments that make a persona useful
  4. Group customers by why they buy
  5. Mark what the evidence can't support
  6. Put the personas to work in messaging and campaigns
  7. Keep personas current as customers change

Start from the customers who bought

Kite builds personas from people who actually bought, because they're the ones whose reasons you want to repeat. With HubSpot connected, Kite reads your won deals, the contacts and companies attached to them, and the notes and meetings your team logged along the way. That shows it who was involved in each purchase and how long each deal took.

Kite separates the roles inside each deal. The person who signs the contract and the person who uses your product every day usually want different things. A finance lead wants to know the product is worth the money and safe to approve. The person using it wants to know it will make their week easier. Kite treats them as separate buyers when the evidence shows they care about different things.

Kite also reads the public LinkedIn profiles of the people on those deals. Their current roles, career histories, and recent posts show what kind of person tends to buy and what they talk about at work. If eight of your last ten buyers were promoted into their role within the past year, that's a pattern worth knowing, because a newly promoted manager has something to prove.

If you don't use a CRM, you can name your best customers in the conversation or share a list. Kite works from a short list and tells you how much weight it can carry.

Gather what customers actually said

Kite reads the places where your customers speak in their own words. Each source shows a different part of the buyer:

  • Sales calls. With Fathom connected, Kite reads the transcripts of your recorded calls. Each line carries the speaker's name, so Kite can tell a prospect's words apart from your rep's. It narrows its reading to calls with people outside your company, so internal meetings stay out of the evidence.
  • Chat and support conversations. With Intercom connected, Kite searches and reads your customer conversations, along with the tags and segments your team already uses. Questions people ask before they buy show what they're unsure about. Complaints after they buy show where the product fell short of what they expected.
  • Your CRM. In HubSpot, Kite reads the notes your team wrote on contacts and deals, plus any feedback survey responses stored there.
  • Interviews and research you already have. You can share interview notes, survey exports, or past research documents in the conversation, and Kite treats them as evidence alongside everything else.
  • Public reviews and discussions. Kite searches the review sites and communities where your buyers write about products like yours, and reads what they say about you.
  • Your competitors' reviewers. Kite reads what your competitors' customers say too. Their complaints show the objections your buyers bring to the whole category, and their praise shows the outcomes the market already values.

Kite favors evidence that shows what people did, such as buying, switching, cancelling, or recommending you, over opinions on their own. It also favors your own buyers over bystanders who have never paid for anything in your category.

Pull out the moments that make a persona useful

Kite reads each source for six things: the trigger that made someone start looking, the cost of the problem, what they tried first, the objection that almost stopped them, the outcome in their words, and any phrase that sounds like a headline. Those six things are what a writer needs to make a page or an ad land.

Quotes stay word for word, with the call, conversation, or review they came from and its date. Kite doesn't tidy a customer's sentence into marketing language, because the customer's version is usually the stronger one. A prospect who says "I'm rebuilding the same spreadsheet every Monday" has given you a better line than "unified reporting."

Kite then ranks what it found by how often it comes up and how strongly people say it. A complaint that appears in thirty conversations earns a place even if each mention is mild. A rare phrase said with real force can earn one too, because it might be the hook for your next campaign. Kite says which kind of evidence each finding rests on, so you can tell a common theme from a single memorable quote.

Group customers by why they buy

Kite groups customers into personas by the reasons they buy, not by age, gender, or job title alone. Two people with different titles who started looking for the same reason belong in the same persona. Two people with the same title can need different personas if one is a founder at a ten-person company and the other runs a department at a large one.

Most businesses end up with two to four personas. Kite proposes the number the evidence supports, and it names each persona by the situation the person is in rather than inventing a first name and a backstory.

Here is how one persona might read for a project-management tool sold to architecture firms, built from forty sales calls and two hundred support conversations:

  • The project architect running three jobs at once. A senior architect who manages projects day to day and uses the tool every day. They don't sign the contract, but they usually find the product and bring it to the principal who does.
  • What made them start looking. A missed permit or consultant deadline that cost the firm money or a client's trust. On one call a prospect said, "We missed the planning submission because the date lived in someone's inbox."
  • What it costs them. Evenings spent chasing consultants for updates, and the worry of being the person blamed when a date slips.
  • What they tried first. A shared spreadsheet and a general-purpose task app. They gave up on both because nobody outside the firm would update them.
  • What almost stopped them. The fear that consultants wouldn't log in. This came up in more than half the sales calls that stalled.
  • The outcome they describe. "I stopped being the human reminder system."
  • Evidence and gaps. Twelve quotes from calls and support conversations, each linked to its source. Kite hasn't heard from anyone at a firm with more than fifty people, so it marks the persona as untested for larger firms.

Mark what the evidence can't support

Kite labels every line of a persona by where it came from. A line is either a quote Kite read at its source, a fact from an export or document your team supplied, or an inference Kite drew from several sources. A quote is never presented as an inference, and an inference is never presented as a quote.

When a part of the persona has no evidence behind it, Kite marks it as unknown and doesn't fill it in from general knowledge of your industry. That's what keeps a persona honest. A gap you can see is far more useful than a confident guess, because you know exactly what to go and find out.

Every persona ends with a gap list. It names what's still unknown, where the evidence is thin, and the cheapest way to close each gap. That might be three more customer calls with a particular kind of buyer, a short survey to existing customers, or an export of last quarter's cancellation reasons.

Put the personas to work in messaging and campaigns

Kite saves the agreed personas to your team's wiki, and its writers read them before they write an ad, a landing page, or a campaign email. That's the difference between a persona that gets used and one that sits in a slide deck. When Kite writes a page for the project architect above, it opens on the missed deadline and answers the worry about consultants logging in, because the persona says that's what this buyer needs to hear.

Personas feed directly into the rest of your messaging. Kite uses them to build a messaging framework for each buyer, which sets the order in which each persona should hear your messages and the proof that convinces them. They also shape the outbound emails Kite writes, the ad copy it prepares for Google, Meta, LinkedIn, TikTok, and X, and the landing pages it builds for each campaign.

Kite can carry the personas into your tools as well. With HubSpot connected, it can record each contact's persona on the contact record after you approve how contacts map to personas, so your team can build lists and campaigns for one persona at a time. With LinkedIn Ads connected, it can look up the targeting options that match a persona and prepare draft campaigns for your team to launch.

If your team needs the personas outside of Kite, Kite can publish them as a report with a shared link, with the quotes and sources included, so sales and product can read the same version you approved.

Keep personas current as customers change

Personas go stale when the market moves and nobody rereads the evidence. Kite can reread your new calls, conversations, and reviews on a schedule, such as once a month, and bring you any changes worth making.

Kite also watches how the work built from each persona performs. It reports replies to the emails it sends and, on a website it hosts, conversions on the pages it builds for each campaign. If messages written for one persona keep going quiet, or a new kind of buyer starts showing up on your calls, Kite brings you the evidence along with a proposed change to the personas.

What you get

Two to four buyer personas

Each one describes a group of real customers by their role in the purchase, their trigger, the cost of the problem, what they tried first, their objections, the outcome they want, and the phrases they use.

A source for every claim

Each quote links to the call, conversation, review, or document it came from, and every inference is labeled as one.

A phrase bank

The exact words your customers use for the problem and the result, grouped by persona, ready for headlines, ads, and email subject lines.

A gap list

What the evidence doesn't yet show for each persona, and the quickest way to find out.

Personas that stay in use

Saved to your team's wiki so every ad, page, and email Kite writes starts from them, and shareable as a report for the rest of your team.

Updates over time

Proposed changes to the personas as new calls, conversations, and campaign results come in.

How to get started

Start a conversation with Kite in the Kite web app or in Slack, and ask in plain words. You might say, "Build buyer personas from our sales calls," or "Who are our target customers, and what do they care about?"

Kite starts from what it learned about your business when you signed up and tells you which evidence it would read. Connect Fathom if you record your calls there, Intercom if your customers talk to you in chat, and HubSpot if your deals and contacts live there. Kite's LinkedIn data and public web research are built in, so they need no connection. You can also share interview notes, survey results, or a list of your best customers directly in the conversation.

Kite reads the evidence and sends you draft personas with their sources and gaps. You push back on anything that doesn't match what you know, and Kite revises them with you. Once you agree, Kite saves them and uses them in the work it ships. If you'd like Kite to keep them current, ask it to reread new evidence on a schedule.

Integrations used to build buyer personas

Kite's built-in research handles public reviews, community discussions, and LinkedIn profiles without any connection. Connecting the tools where your customers talk to you gives Kite the evidence that matters most: your own buyers, in their own words.

Find out who bought and who they are

HubSpot logo
HubSpot

Kite reads your won deals, the contacts and companies on them, and the notes your team logged, then records each contact's persona back on the contact once you approve the mapping.

LinkedIn (data)

LinkedIn data is built into Kite. It reads the public profiles and recent posts of the people who bought from you, which shows their roles, career paths, and what they talk about at work.

Common questions

What are the four types of buyer personas?

Most articles that list four types of buyer personas sort people by how they make decisions. Competitive buyers want proof they'll come out ahead, spontaneous buyers decide quickly and want the short version, methodical buyers want detail and data, and humanistic buyers care about the people and relationships involved. Other lists sort by role in the purchase instead: the decision maker, the champion who brings the product in, the daily user, and the influencer who weighs in. Both lists are useful prompts, but neither tells you who your buyers are. Kite starts from your evidence and lets the groups come from it, and it often finds that a decision-making style or a buying role is one trait inside a persona rather than the persona itself.

Can I build buyer personas with ChatGPT or a free AI persona generator?

You can, but the result depends entirely on what you give it. A persona generator or a general chatbot asked to "create a buyer persona for my product" writes a plausible profile from general knowledge of your category, so it describes an average customer rather than yours. If you paste in real transcripts and reviews, a chatbot can find patterns in them, but you have to collect the evidence, keep track of sources, and repeat the work when things change. Kite gathers the evidence from your connected tools and public sources itself, keeps every quote tied to its source, and uses the finished personas in the work it ships.

How many buyer personas should I have?

Most businesses need two to four. One persona usually blends people who buy for different reasons, and the messaging ends up speaking to none of them. More than four becomes hard for a small team to write for, and some of the groups usually turn out to want the same thing. Kite proposes the number your evidence supports and explains why each persona is separate from the others.

What's the difference between a buyer persona and an ideal customer profile?

An ideal customer profile describes the companies you should sell to, using traits like industry, size, and funding. A buyer persona describes the people inside those companies, or the consumers you sell to directly, and what makes them act. The profile tells you which accounts to pursue, and the personas tell you what to say once you reach the people there. Kite builds both, and its ideal customer profile work includes the titles and seniority of the people who buy, which the personas then describe in depth.

Can Kite build personas if I don't record calls or have many customers?

Yes. Kite can work from public reviews of your product and your competitors, community discussions where your buyers write, your CRM notes, and whatever interviews or survey results you share. With less evidence, more of each persona is marked as inference or left as a gap, and Kite tells you which sources would strengthen it most. Even a handful of recorded customer conversations usually adds more than any amount of general research, so the gap list often starts there.

Questions this page answers

  • How do I build buyer personas with AI?
  • Can an AI marketing agent build buyer personas?
  • How do I create buyer personas from real customer data with AI?
  • AI agent that builds personas from interviews, reviews, and CRM data
  • How to build personas that aren't made up with AI
  • Can AI turn customer conversations into buyer personas?
  • How does an AI marketing agent ground a persona in evidence?
  • targeted customers

About the author

Devon Wells leads product marketing at Kite, the AI marketer. He spent a decade marketing industrial AI, from self-driving trucks to computer vision, but now Kite handles most of the work itself.

Tell me your goal and I’ll get to work.

Get started for free