How to Use Sales Call Transcripts to Generate AI Content

Ask any seasoned content professional for tips on creating persuasive content, and they’ll point you to the same resource: sales calls.

Sales call transcripts can offer a brand new perspective on your product or the customer’s pain points. And unlike traditional keyword research, which surfaces topics that are already gaining traction, sales transcripts can point you to emerging trends. 

When you’re the first to cover a topic, you’ll grab the attention of your customer base, show them you understand their problems, and grow brand authority.

Businesses like Copy.ai have already used AI to turn their sales transcripts into content drafts, cutting back on production time while scaling content creation.

To understand how the team at Copy.ai used sales call transcripts for AI content, Tim Soulo (Chief Marketing Officer at Ahrefs) spoke to Zac Harris (founder of SEO and AI agency Rankd. and former Head of Demand Generation at Copy.ai). We outline their step-by-step process below.

1. Record sales calls

Recording and transcribing sales calls is the first step to turning them into AI content. Avoid relying on meeting notes – AI needs as much context as possible.

A tool like Gong is a great option. It records your Zoom calls and automatically uploads the recording and transcript to your account when the meeting is done.

Then, you’ll use this transcript to generate AI content, such as a LinkedIn post or a blog article.

2. Build an automated workflow

An automated workflow is a more efficient choice for creating AI content, especially if the sales team takes several meetings per day. 

In the case of Copy.ai, the team built an automated workflow with the help of a solutions engineer. It would trigger every time Gong generated a sales call transcript and send it to a Copy.ai workflow in Airtable, which would generate the first draft of AI content. Then, it pasted the draft into a Notion database for a human writer to review and improve.

You could still do this manually if you don’t have the resources to automate AI content writing. 

As an example, let’s use ChatGPT with the following prompt: “Pull the key points from this transcript. Generate LinkedIn post ideas for each key point.” (Customize the prompt with instructions on style, tone of voice, etc.)

Paste the sales call transcript (or a specific section) and wait for the AI content to generate.

Here are the results using a section of the interview transcript between Tim and Zac:

3. Bring a human into the loop

Before you hit “Publish”, use a content editor or writer to review the results, remove inaccurate information, and tweak the tone.

For instance, Copy.ai used freelance writers to check if AI included any hallucinations about their product or a specific topic in the first draft.

Where appropriate, add visual elements to make the content more engaging. For example, a long blog post will benefit from a diagram breaking down key information or workflow screenshots.

Potential issues to keep in mind

When generating AI content from sales transcripts, the first issue you might run into is the large volume of content. 

If the sales team does an average of five calls per day, this means generating 25 drafts per week, which can quickly snowball into thousands of drafts. This is precisely what happened at Copy.ai where the team had 3,000 AI-generated articles at one point. It’s too overwhelming for the average content team to handle.

In addition, you might find that many of these drafts are similar – especially if the salespeople talk to the same kind of customer. In this case, it’s a better use of your resources to analyze sales call transcripts with AI and pull out unique topics for further ideation or customer engagement.

“We had an AI workflow that was built out for our lifecycle manager. It would go through pretty much the same process using the transcript, but then it would pull out the customer’s pain points, any objections, what types of assets should we send them post-call, and things like that to help the manager build out sequences,“ explains Zac.

Experiment for the best results

AI content generation is still in its early stages, which means that one approach might not work for all content teams. 

If you’re unhappy with your AI content drafts from sales call transcripts, brainstorm a different strategy. For example, you could use AI to summarize sales calls and send the summaries to the content team for topic research.

Over time, these experiments will help you build an AI-supported content engine that fuels your blog, email newsletter, and many other channels.