Turn customer interviews into findings

Turn a stack of interview notes and sales call recordings into findings about who your best customers are, what pushed them to buy, and what nearly stopped them, with the customer's own words behind every finding.

Key Takeaways

  1. Useful interview findings answer a short list of practical questions: what happened right before the customer went looking, what the problem was costing them, what they tried first, what almost stopped them from buying, and what they say they got. Each finding carries the customer's exact words.
  2. Kite reads the transcripts and notes you already have, whether they live in Fathom, Google Docs, or Google Drive or you paste them into a conversation. It separates what customers said from what your interviewer said.
  3. Kite ranks themes by how often they come up and how strongly people say them. It keeps every quote word for word with its source, and it labels its own interpretations so you can always tell a quote from a conclusion.
  4. The findings go into your team's knowledge base, where Kite reads them before it writes an ad, a landing page, or a campaign email. Kite then proposes the first changes the findings point to.
  5. Kite also lists what the interviews didn't answer and the next conversations worth having.
On this page

How do you turn customer interviews into findings?

To turn customer interviews into findings, read every transcript with the same few questions in mind, pull the customer's exact words for each answer, and count how many interviews say the same thing before you call it a pattern. The questions that matter most for marketing are:

  • What happened right before they started looking for a product like yours?
  • What was the problem costing them, in hours, money, or risk?
  • What did they try first, and why didn't it work?
  • What almost stopped them from buying?
  • What do they say they got, in their own words?

Most teams never get that far. Interviews and sales calls pile up faster than anyone can read them. A 45-minute call runs to about 6,000 words, so a dozen interviews is roughly the length of a novel. When someone does find an afternoon, the result is often a summary that turns what customers said into marketing language, such as "users want better visibility." That line can't be tested, and it can't be used in an ad. The phrase a customer actually used, such as "I was rebuilding the same spreadsheet every Monday," can.

An AI agent is well suited to this work because most of it is careful reading, done the same way across every transcript. It doesn't get tired by the ninth interview, and it doesn't drift toward the themes it noticed first. The harder part is making sure the findings reach the copy your team writes next, which is where an agent that also does the writing has an advantage.

Kite is an AI marketer. It researches your business, builds a growth strategy, and ships the work that brings you new customers. You can work with Kite in your browser or in Slack. Here's how it turns your interviews into findings and puts them to use.

How Kite turns customer interviews into findings

Kite starts with what it learned about your company when you signed up: your positioning, your products, your customers, and your competitors. That context tells it what to listen for. Then it gathers the interviews and calls you point it to, or finds them itself in your connected accounts.

Kite reads each conversation and pulls out the customer's own lines. It sorts what it finds into the triggers that sent customers looking, the objections that nearly stopped them, the alternatives they tried, and the outcomes they describe. It ranks those themes, keeps every quote word for word with its source, and writes the findings into your team's knowledge base. From there, Kite proposes what to change first, such as a homepage headline, a page that answers a common objection, or an update to your ideal customer profile. It also lists what the interviews left unanswered.

  1. Gather your interview notes and sales call recordings
  2. Separate what customers said from what your team said
  3. How do you find buying triggers in customer interviews?
  4. Find the objections that nearly stopped the sale
  5. Decide which findings are patterns and which are one-offs
  6. Save the findings where your marketing will use them
  7. Turn the findings into changes to your marketing
  8. Keep the findings current as new calls come in

Gather your interview notes and sales call recordings

Kite works from the conversations you already have, wherever they're kept. You don't need to clean them up or move them into one place first.

If you record calls with Fathom, connect your Fathom account and Kite reads the transcripts and summaries directly. It can narrow the calls to ones with people outside your company, to a particular date range, to calls with one customer's email domain, or to calls one teammate recorded. That's how a request like "What did prospects push back on in last month's demo calls?" turns into a reading list without anyone picking calls by hand. Kite scans Fathom's summaries first to find the conversations that matter, then reads the full transcripts of those.

If your interview notes and transcripts live in Google Drive or Google Docs, Kite searches for them by name, folder, date, or the words inside them, and reads what it finds. You can also paste a transcript into your conversation with Kite, upload transcripts as PDFs, or share the files in a Slack channel Kite has joined.

Separate what customers said from what your team said

A transcript is only useful as evidence when you know who said each line. Fathom transcripts and most notetaker exports label every line with the speaker and a timestamp, so Kite can set your interviewer's questions and your sales rep's pitch aside and work from the customer's words alone.

This matters more than it sounds. On a sales call, your rep may describe the product's benefits several times, and a careless summary can mistake the rep's pitch for the buyer's opinion. Kite keeps the two apart.

Notes need different handling. When a teammate writes "she said reporting takes forever," that's the teammate's account of the conversation, not a quote. Kite treats it as a summary unless the notes mark it as the customer's exact words, so a paraphrase never ends up in your findings presented as something a customer said.

How do you find buying triggers in customer interviews?

You find buying triggers by looking for what changed in the customer's world just before they started looking, not for why they say they liked your product. A trigger is usually a specific event: a new hire, a lost client, an audit, a price increase from their old vendor, or a month that went badly. Customers rarely call it a trigger. It shows up in sentences like "after we lost the Henderson account" or "when our second designer quit."

Kite reads each transcript for those moments and records each trigger with the customer's quote and the call it came from. Alongside the trigger, it notes what the problem was costing the customer in practice, such as hours each week or a client at risk, and what it cost them personally where they say so. It also records what they tried first and the reason they gave for moving on, because those are the alternatives your marketing will be compared with.

Triggers are the most useful finding for timing your marketing. If six of ten customers started looking right after hiring their first marketer, that event tells you when to reach new prospects and what to say when you do.

Find the objections that nearly stopped the sale

Objections tell you what your website and sales process have to answer before a buyer will commit. Kite looks for the moments a customer hesitated or pushed back, and the moments they describe hesitating before they bought.

It groups those objections into the kinds buyers usually raise: price, trust, the effort of switching, whether the product fits their situation, and timing. Then it ranks them by how often they come up and how strongly they're said. A worry that appears in most interviews earns a place in the findings even when each mention is mild. A single objection said with real force can matter too, and Kite says which kind each one is.

Sales calls are especially good for this, because buyers ask their questions out loud. "Do we need an engineer to set this up?" asked on four separate demo calls is a gap on your website, and Kite treats it as one.

Decide which findings are patterns and which are one-offs

A finding is only as good as the evidence behind it, so Kite shows the evidence for each one. Every theme is marked as recurring or seen once. Kite gives a count, such as "five of eight interviews," only when it has actually counted the conversations behind it, and it doesn't turn a small sample into an impressive-sounding percentage.

Every quote stays exactly as the customer said it, with the call or document it came from and the date. Kite doesn't merge two quotes into one or tidy a quote into smoother language. When Kite draws a conclusion from several interviews together, it labels that as its own reading rather than presenting it as something a customer said. That way you can check any finding against the source in a minute, and you can tell at a glance which findings rest on a customer's words and which rest on Kite's judgment.

Kite also keeps a phrase bank: the exact lines that sound like a headline or a hook, grouped by the feeling behind them, such as relief, frustration, or dread. These are the lines your writers will want most.

Save the findings where your marketing will use them

Findings that sit in a slide deck stop shaping your marketing within a month. Kite writes them into your team's knowledge base as a customer voice page, the same place it keeps your positioning, your brand voice, and your research on competitors.

Kite's writers read that page before they write an ad, a landing page, or a campaign email. A headline Kite drafts next quarter can use the phrase a customer said on a call this month, and the source of that phrase is one click away. When the interviews change what you know about your best customers, Kite proposes updates to your ideal customer profile as well.

If your team wants something to read and share, Kite can also write the findings as a Google Doc with the quotes behind each one, or as a report with charts and tables at a shareable link.

Turn the findings into changes to your marketing

Kite doesn't stop at the findings. It proposes what to change first, based on which findings are strongest and where your marketing currently says something different from what customers say.

The changes usually fall into a few groups:

  • Your messaging. Kite rewrites a homepage headline, a pricing page section, or the opening line of an outbound email in the words customers used. When different kinds of buyers described different triggers, it can build a messaging framework for each buyer.
  • Pages that answer objections. When the same objection shows up across many conversations, Kite drafts an FAQ answer, a new section, or a page that answers the question before the next buyer has to ask it.
  • Customer proof. When a customer describes a specific result, Kite flags them as a candidate for a case study and drafts the request asking for their permission.

Kite asks for your approval before it publishes a page or sends an email. On a website Kite hosts, it can test a new headline against the old one and keep whichever brings in more signups. On a site measured elsewhere, it compares performance before and after the change through your connected analytics.

Keep the findings current as new calls come in

Customer research gets stale as your product, your market, and your customers change. Kite can keep the findings current without anyone remembering to ask.

Set up a weekly job, and Kite reads the new external calls in Fathom or the new files in your "Customer calls" folder in Google Drive since its last check. It adds new quotes to the existing themes, marks a theme as recurring once it shows up again, and tells you when something new appears, such as a competitor customers have started naming.

Kite also keeps a list of gaps: the questions the interviews haven't answered yet, where the evidence is thin, and the cheapest way to close each gap, such as three more calls with customers who nearly churned. When you're ready for that next round of conversations, Kite can write the interview guide aimed at those gaps.

What you get

A customer voice page

Triggers, costs, failed alternatives, objections, and outcomes, each with the customer's exact words and the source, saved in your team's knowledge base for every later piece of work.

Ranked objections

The concerns that nearly stopped buyers, ordered by how often and how strongly they came up, with each one marked as recurring or seen once.

A phrase bank

The customer lines most worth using in headlines, ads, and emails, grouped by the feeling behind them.

A shareable findings doc

The same findings as a Google Doc or a report at a link, with the quotes behind each one, for your team or your product lead.

Proposed changes

The headline, page, profile, and case study work the findings point to, ready for your approval.

A gap list

What the interviews didn't answer and the next conversations that would answer it.

How to get started

Get started for free and tell Kite what you want to learn, in plain language. You might say, "Read our last ten customer interviews and tell me why people buy," or "What objections came up on our demo calls this month?"

Then give Kite the conversations. Connect Fathom if you record calls there, or connect Google Drive or Google Docs if that's where your notes and transcripts live. Kite sends a connection link the first time a task needs one. You can also paste transcripts into the conversation or upload them as PDFs, with no connection at all.

Kite reads the conversations, writes the findings, and brings you the themes it found and the changes it would make first. Correct anything it got wrong, and Kite updates the findings. Kite asks for your approval before it publishes a page, sends an email, or asks a customer for permission to quote them.

Integrations used to turn customer interviews into findings

Kite can work from transcripts you paste or upload, so you don't need to connect anything to start. Connections let Kite find the conversations on its own, keep reading new ones as they arrive, and deliver the findings where your team already works.

Common questions

Can ChatGPT do a thematic analysis of interview transcripts?

Yes. If you paste transcripts into ChatGPT and ask for themes, it will group what customers said and suggest patterns. The work you're left with is what happens next. You have to check that the quotes are exact, keep the findings somewhere your team will find them, and carry them into the next piece of copy yourself. Kite keeps each quote word for word with its source, saves the findings in your team's knowledge base, reads them before it writes your marketing, and proposes the changes they point to.

Does Kite join or record customer interviews?

No. Your interviewer runs the conversation, and a notetaker such as Fathom records and transcribes it as usual. Kite reads the transcript afterward. Before the interview, Kite can help you prepare by writing an interview guide built around the questions your findings haven't answered yet.

What if I only have audio recordings, not transcripts?

Send a transcript whenever you can. Kite transcribes an audio file you send it, but only the first ten minutes, which suits a voice memo better than a full interview. Zoom, Google Meet, Microsoft Teams, and most notetakers can export a full transcript, and Kite reads Fathom transcripts directly once Fathom is connected.

How many interviews do you need before the findings mean anything?

You can learn something useful from five or six interviews with the right customers, as long as you're honest about how strong each finding is. Kite marks every theme as recurring or seen once and gives counts only when it has counted the conversations behind them. As you add interviews, findings that were seen once either repeat or fade, and Kite updates the page to match.

What is the 80/20 rule in customer interviews?

The 80/20 rule says the customer should do about 80 percent of the talking and the interviewer about 20 percent. It matters for analysis because the findings come from the customer's lines. Because transcripts label each speaker, Kite can point out interviews where the interviewer did most of the talking, so you know which conversations carry the least evidence.

Can we use quotes from interviews in our marketing?

Yes, with the customer's permission. Kite uses interview quotes freely inside your team's knowledge base and in drafts for your review. Before a quote appears on your website or in a case study, Kite drafts a request asking the customer to approve the exact wording, and it asks you to approve the page before it publishes.

Questions this page answers

  • How do I turn customer interviews into findings with AI?
  • Can an AI marketing agent turn customer interviews into findings?
  • How do I analyze customer interview transcripts?
  • How do I turn sales call recordings into insights?
  • How do I find buying triggers in customer interviews?
  • How do I synthesize customer research notes?

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