Week 3: Unlock Customer Insights from Sales Calls

The Experiment
Watching hours of sales call recordings and taking notes can be soul-crushing. It’s useful, but it doesn’t scale and after the 5th call, your brain feels like a damp sponge. I wanted a better way to extract the nuggets from these calls: the customer’s major pain points, purchase drivers, and why this solution is transformative for their business.

I wanted to find a process that captures key insights without losing the nuance or spending days combing through transcripts.

The Process
Here’s what I did:

1. Upload Sales Calls to Otter.ai: First, I converted the spoken conversations into text. Otter.ai is great for this—fast and accurate enough to get a clean transcript of every call (it handles different accents remarkably well)

2. Focus with Notebook LLM: I uploaded the transcripts into Notebook LLM, an AI tool that only works with the data you feed it. It’s a better tool for this use case because it isn’t going to add other information, its only dealing with what you upload to it; no hallucinations —just my data.

3. Interrogate the Data: Using Notebook LLM, I was able to dig deep. I could ask for specific insights, like:

   • What were the main pain points clients mentioned? (make sure in your prompt you’re only getting client feedback, not sales/account manager discussions)

   • Why did they think our product would transform their organization?

   • Can you pull direct quotes from clients to illustrate these points?

4. Validate and Refine: I always double-checked the AI’s “homework” by going back to the transcript to ensure quotes were accurate and the nuance wasn’t lost.

The Outcome
This process worked well. Notebook LLM is way faster and less draining than doing it manually. It’s especially powerful for capturing the client’s voice—real quotes that you can weave into presentations or reports.

Bonus Feature: Notebook LLM can even create a podcast where it talks through the insights from your data. It’s a fun touch if you want to share a summary with your team or client without making them read a wall of text.

There’s some prep work upfront—cleaning and uploading transcripts takes a little time—but once it’s set up the analysis is fast.

Key Takeaway
If you’re dealing with loads of customer or sales data, pairing Otter.ai with Notebook LLM (you can add up to 50 documents, links or videos) is highly recommended. You can quickly pinpoint what matters most, get customer quotes to back up your findings, and save hours of time.

Pro Tips for Beginners:

1. Use Thematic Prompts for Notebook LLM: Instead of vague queries, create thematic prompts like:

• "Summarize all mentions of budget concerns."

• "Highlight examples where the customer described their ideal solution."
  This approach ensures your insights are aligned with specific goals.

2. Ask Specific Questions: The more focused your queries in Notebook LLM, the better your results. Don’t just ask “What are the pain points?”—get granular and iterate based on what has been said previously

3. Validate Everything: AI is great, but double-checking for nuance ensures accuracy (and saves you embarrassment).

Want to Try It Yourself?

• Use Otter.ai for speech-to-text.

• Upload transcripts to Notebook LLM for hyper-focused analysis.

• Pull real client quotes to make your findings more compelling.