Home · Business · Marketing & Content · Market research

YES

As of 13 August 2026, AI can summarise your customer interviews.

This still needs a person who signs their name to it.

Can you do it?

5 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ityou

What the alternative costsThe supplied tool data gives no price for a human interview-summary service.

If this goes wrong: the summary hides a minority view or turns a tentative comment into a firm finding, and your product or marketing decision follows it.

What to actually do

  1. Hand it to a person

    The route this page recommends

    A person who owns the outcome does this end to end, worth it when the failure is dear.

  2. Use a tool built for this

    Second choice

  3. Do it yourself

    The distant third

    A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

    How to actually do it

    1. Open the interview recordings or transcripts and remove names, contact details, company secrets and any other information that identifies a customer.
    2. Confirm that each participant agreed to the recording and the intended research use, then label each interview with a neutral identifier.
    3. Gather the research question, interview guide, relevant product context and the cleaned transcripts or notes in one document.
    4. Paste the prompt and the material into a chatbot, keeping each interview clearly separated by its identifier.
    5. Ask the chatbot to produce the summary, themes, dissenting views, exact quotes and weak findings in the requested sections.
    6. Compare every theme, count and quote in the draft against the relevant transcript, and correct anything missing, overstated or paraphrased as verbatim.
    7. Send the checked summary to the research or product team with the source transcripts or notes attached, rather than presenting the AI summary as the evidence itself.

    Prompt

    Summarise the customer interviews below for a UK market-research team. Use only the supplied material and do not invent facts, motives, quotes or levels of confidence. Remove or generalise names, contact details and other identifying information. Produce: 1) a short executive summary, 2) the main themes with the number of interviews supporting each theme only when that number can be established from the material, 3) important differences or dissenting views, 4) customer needs and problems, 5) notable verbatim quotes copied exactly and labelled with the interview identifier, 6) unanswered questions and possible follow-up research, and 7) findings that are too weak or ambiguous to rely on. Separate what customers explicitly said from your interpretation. Flag every claim that needs checking against the source. Do not recommend a product, campaign or business decision unless I ask for that separately.

Research question: [insert research question]
Interview identifiers: [insert non-identifying labels]
Interview guide or context: [paste here]
Interview transcripts or approved notes: [paste here]

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

What it gets wrong

  • AI cannot decide whether a repeated complaint is important because of your market, strategy or customer mix.
  • AI compresses hesitation, humour and context, so a seemingly clear theme can lose the meaning carried by the conversation.
  • AI can miss a minority view or treat an interviewer prompt as customer demand.
  • AI cannot establish that a quote is representative or that the interviews support a business decision.
  • Uploading recordings or transcripts can expose private customer information unless you have consent and suitable handling controls.

Even on a YES, the friction has a name: judgement under ambiguity, consent and privacy and verification cost.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability1
Effort delta2
Total8 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI summarise recorded customer interviews?
Yes. It can transcribe or process the recordings and produce themes, quotes, differences and action points, but you need to check those against the source before using them.
Can ChatGPT analyse customer interview transcripts?
Yes, if you provide the transcripts and clear research instructions. Remove identifying information first, and separate what customers said from the model's interpretation.
How accurate are AI summaries of customer interviews?
They are useful as a first pass but can omit minority views, flatten uncertainty or present an interpretation as a finding. Check each important theme, count and quote against the transcript or recording.
Is it safe to upload customer interviews to AI?
Only when your consent, privacy and organisational data-handling requirements allow it. Remove identifying details where possible, avoid uploading unnecessary sensitive information and check the tool's terms before use.

Nearby answers

All Market research rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.