How to Summarize Sales Calls and Analyze Competitors with AI
According to sales SaaS provider ServiceBell, the average B2B salesperson makes 35 calls a day, which adds up to 700 calls per month, or 8,400 calls per year.
This number alone is a solid reason to start considering solutions that help salespeople manage their workload more efficiently, such as AI summarizers.
But what are AI summarizers?
Also referred to as “automatic text summarization tools”, AI summarizers are solutions that use artificial intelligence to generate an abstract of a larger body of text.
In short, an AI summarizer captures the most important points within a piece of content and presents them in a concise way that preserves the original meaning of said content.
AI summarizers can be used to condense various types of content, including articles, reports, essays, and even transcripts of spoken conversations like - you guessed right - sales calls.
In this tutorial, we’ll show you how to create your own AI-powered sales call summarizer with market-leading tools like Gong, ChatGPT, and Slack.
Before we jump into details, let’s address a number of important details, such as the tools you’ll need to build an AI summarizer.
Calling better shots with AI and Make
On top of producing a summary of every sales call made by your team, this solution will also:
Get the competitors of your company/product.
Identify strengths and weaknesses in the competitors’ offerings.
Send this information to your sales team over Slack.
To build it, you’ll need accounts for the following tools: Make, Gong, OpenAI, and Slack.
If you’re new to Make, we recommend following the tutorial below, but it’s worth noting that the solution is also available as a pre-configured Make template that you can use right away:
Have everything ready? Then let’s build our own AI summarizer from scratch.
Step 1: Create a new Make scenario and add the Gong app
Once your Make account is ready, create a new scenario from your dashboard.
Next, click on the purple circle, search for the Gong app and select the Watch new calls module, which will act as the trigger for this scenario.
Now it’s time to configure the Gong module.
Start by connecting your Gong account to Make selecting Gong Basic as the Connection Type.
You can click on the ‘?’ symbol to open the help page if you need help finding your Access Key and Access Key Secret in your Gong account.
After establishing the connection, you’ll need to specify how many sales calls Make will retrieve each time the scenario executes.
Specify the number in the Limit field, and then click OK to finish configuring this module.
Step 2: Add the second Gong module
To add the second module in the scenario, click on the plus sign on the side of the first module.
We want to fetch the transcripts of the calls, but since there isn’t a dedicated Gong module in Make to do this, we need to use the Make an API call module.
To configure it, start by pasting the following path in the URL field: /v2/calls/transcript In the Method field, select POST.
Then, copy and paste this JSON structure in the Body field:
{
"filter": {
"fromDateTime": "{{addYears(now; -1)}}",
"toDateTime": "{{now}}",
"callIds": [
"{id}"
]
}
}
You’ll need to add a couple of data operators to the JSON structure - please use the image below as a reference, and make sure it includes the Call ID data item!
The date range is set up so that all calls for the past 24 hours are fetched, and the scenario will execute at 8 AM every morning.
You can adjust the date range in the JSON structure and the execution schedule in the scenario if you need to.
When the transcript is received via this module, we need to iterate through the Sentences[] array to output each sentence of the transcript.
Important note: If we don’t, we will only receive the first sentence and not the entire transcript!
Once you do this, click OK to finish configuring the module.
Now, we need an output of data before we move on to the next module.
To get the data, execute the scenario: Right-click on the trigger module, then click Choose where to start, select the Choose manually option, and select a call.
Next, click OK and then click the Run once button to execute the scenario.
Step 3: Add an iterator and a text aggregator
At the bottom of the scenario builder, select the Iterator and attach it after the second Gong module.
Configuring the iterator is easy. Just map the Sentences[] array to the field as shown below.
As it stands, the iterator outputs each sentence one by one, but there’s a catch: We can’t send the sentences one at a time to ChatGPT.
Before doing so, we need to aggregate the sentences into a single text first.
To achieve this, we’ll use the Text Aggregator (also found at the bottom of the builder).
To configure the Text Aggregator module, select Iterator as the Source Module and then map the Text data item to the Text field.
Once done, click OK. Now we can send the full transcript of our sales calls to ChatGPT automatically!
Step 4: Add the OpenAI app
Here we need to add another module to our scenario. Search for the OpenAI app and select the Create a Completion module.
Start by connecting your OpenAI account to Make, and configure the module as follows:
Method: Create a Chat Completion
Model: gpt-3.5-turbo-0301 (currently has the highest limit of 16k tokens)
Message role: User
Message content: Summarize the text '{text}' and give me a list of bullet points with key insights and the most important facts.
Don't forget to map the Text data item from the Text Aggregator to the text as you see below.
Once ready, click OK.
Step 5: Add the Slack module
Finally, add the Slack > Create a message module to the scenario. This module will automatically notify the sales team about newly available call summaries.
To configure the module, start by connecting your account, and then choose the Select from the list option.
After doing this, select the Channel Type and then the Channel/User that will receive the resulting messages.
You can compose your message and map the data that you want from the modules.
Just keep in mind that the output (the call summary) from the OpenAI module is the Choices[].Message.Content data item, which you should map in the message.
Once done, click OK.
Step 6: Schedule and save the scenario
To schedule the scenario, click the clock icon and set your scenario to execute daily at 8 AM. Click OK once ready.
To wrap it up, save the scenario, switch it on and you’re good to go!
With this scenario active, Make will automatically summarize every sales call coming from Gong.
Final thoughts: What are the benefits of using an AI summarizer?
Summarizing sales calls can have numerous benefits for businesses, including:
Record keeping: A summary provides a record of the call, including topics discussed, decisions made, and action items identified. This can be valuable for future reference (and also quite easy to digest when compared to full transcripts).
Customer insights: Call summaries are a source of insight into customer needs, objections, and preferences. This information can be used to adjust and improve future interactions.
Knowledge sharing: When a team member learns something valuable from a call, that insight can be shared easily thanks to the availability of summaries.
Follow-ups: Summaries can act as a reminder for follow-up actions. These could be sending additional information, arranging another call, or moving forward with a deal.
Performance reviews: Call summaries can provide objective data for performance reviews. Managers can use these to understand how well sales reps are doing without having to go through the pain of listening to full recordings.
Analytics: By adding call summaries to your CRM system, you can track how many calls it takes to close a deal, common customer concerns, or the effectiveness of different sales strategies.
It goes without saying that managing hundreds of interactions with potential customers can easily get out of hand, and that’s where AI summarizers shine.
Ready to build yours? Get your Make account, and don’t let the momentum slip away!
