• Find your message in 100,000+ word transcripts in MINUTES, not days
  • For Marketers, Sales teams, Product managers, Customer success, Support teams
  • Takes 5 minutes max (a FRACTION of the time you’d need to do this manually)

AI has been great at automatically transcribing calls for a couple of years. But when it comes to actually looking at all those words and giving you the insights you need?

Not so much.

Until now.

What happens typically is, your interview is over and you have your nicely formatted transcript, including timestamps that you can analyze and look over when you have some time.

So, never.

Instead, you probably end up rewatching the whole interview recording at 5x in hopes of gleaning a few golden nuggets after months.

Let’s put an end to it.

Here’s what to do: Prompt ChatGPT to do the grunt work for you. Have it look through entire interview transcripts and analyze them for you. So you can extract the insights you need to write good copy, without going crazy.

And the best part, you can use this same process to analyze pretty much anything, from interviews (podcasts included), to case studies, sales calls, and even pitch decks.

This is part of what ChatGPT created for me when I used the AI copywriting prompts I’m about to teach you:

You can view the full chat here.

Here’s how to use ChatGPT 3.5 or 4 to get similar results for your business or client.

Step 1: Clean up your VOC data and chunk it down

First, in order to “feed” ChatGPT with your transcript or with any text from other interviews or slide decks, we need to clean it up.

This means we need to make sure we remove all full names and specific references (data privacy is important especially if you signed an NDA), as pasting it into ChatGPT also means they’ll be able to use it to improve their models.

Once you’ve cleaned up your text, it’s time to feed it to ChatGPT. At the time of this writing, we still have a character limit, so we’ll use a quick workaround.

Head over to ChatGPT Splitter and paste your entire transcript/text into this window:

Once done, click on the “Process” button below.

The tool will split the text in as many chunks you need to be able to share it with ChatGPT. Now simply click on “Copy” on the first chunk and paste it in ChatGPT:

Repeat for all chunks one at a time until you’ve copy-pasted all of them and ChatGPT gives you the message “OK: [CHUNK X/X]”.

Step 2: Ask ChatGPT to analyze your text

Next, we’re going to give ChatGPT some context on what frame to use to look at the text (copywriter, marketing expert etc.) and then ask it to analyze our text for us.

In ChatGPT 3.5 or 4, enter the following prompt:


Copy paste this prompt:

You are an expert at sales, positioning strategy and conversion copywriting. Out of the text you just read and memorized above, which come from a customer interview, extract the top key insights that we can use for our positioning strategy and value proposition copy. With each insight, provide the takeaways, actionable steps as bullet points.

Use the following format:

- Takeaway 1: [state the takeaway]
- Action steps for takeaway 1: [state the action steps]

[More takeaways and action steps with the same format if available]

No need to replace anything in this prompt.

In just seconds, ChatGPT will go through your text or interview transcript and extract the insights you need:

I tried this AI prompt for a few different types of assets. Here’s what happened.

I’ve used this prompt or variations of it for any type of marketing or research asset you could think of. The best part is that as long as you have the raw text, you can use ChatGPT to extract any specific insights you think you can gather from it.

For example, I’ve modified my previous prompt to extract more specific takeaways from the interview transcript (objections and surprising facts):

Here’s what ChatGPT gave me:

Again, you can view the full chat here.

Feel free to experiment and change this up as it suits your needs. What matters is that you use good voice of customer data and get going.

FacebookTwitterLinkedInEmail