Functionality Overview

The way this app works is by grabbing the natively generated call transcript from completed calls, and passing them off to your workflow.

This allows you to pass that transcript to ChatGPT in order to generate a call summary, determine the quality of the lead, conditionally add tags, grade the performance of the salesperson, and more.

Here is a video showing how it works to generate a basic call summary:

I’ve listed some example workflows that use this app in order to solve all sorts of business problems.

Which Transcripts App to Choose

You may be confused after seeing two similar apps on this site: Advanced Call Transcripts & Summaries and Advanced Call Transcripts for Workflow.

The main difference is that Advanced Call Transcripts & Summaries uses the built in (native) call transcripts, while Advanced Call Transcripts for Workflow uses an external API to generate a transcript.

This gives Advanced Call Transcripts for Workflow a few distinct advantages:

• This app will work for inbound calls started with an IVR workflow
• This app will transcribe calls in other languages
• This app can redact certain sensitive info such as credit card numbers, social security numbers, etc.
• This app is more accurate with the individual word transcriptions
• Finally, it’s much cheaper to run
• The best possible model is priced at $1.04/hr, while the native transcripts are priced at ~$2.40/hr
• Additionally, you can choose to only transcribe calls that matter to you, saving an additional 10-50% in cost.

On This Page

• ¶ Why Native Transcripts
• ¶ Why Workflow Native
• ¶ Trigger Overview
• ¶ Example Transcript
• ¶ Speaker Talk Times
• ¶ Example Workflows
• ¶ Workflow Snapshots
• ¶ Post-Installation Pages

Native Transcripts

I think at this point I’ve built 3 separate call summary apps for this CRM, with this being the most recent iteration.

The way all of those previous apps worked was to pass the call recording URL off to OpenAI or Deepgram (speech-to-text services) which would generate a transcript.

This would work great in theory, but in practice the quality of the transcription was poor.

This is a problem with most speech-to-text algorithms as you can see from any time spent on YouTube, but the real problem was with speaker recognition.

The algorithm is listening to a single audio channel call recording, and trying to tell the speakers apart based on the sound of their voice.

This often meant that for a two-person conversation, there could be three or four separate speakers listed. With no consistencies between them.

Additionally, we couldn’t determine with any accuracy when the account owner was speaking, versus when the contact was speaking.

Not only does this make formatting difficult, but it means that ChatGPT gets a jumbled mess of speakers, words, and sentences, and has trouble putting together a useful summary for the CRM user.

With native transcripts, those problems are largely solved.

We can say with 100% accuracy who the account owner is vs. who the contact is, which helps ChatGPT know what the CRM user wants to remember.

There are also only two speaker channels (CRM user and contact) since it has a multi-channel recording, so the transcript can be reliably formatted into text.

TL;DR: Native transcripts are better.

Workflow Native

Another issue with my previous call summary apps was the difficulty in adjusting the prompt.

In fact, it’s more than the prompt since it’s only with this iteration that I realized if you just had the transcript, you could do anything from the example workflows, and more.

With this app, since the transcripts are delivered via a workflow trigger, you can do whatever you want with it.

Generate a simple summary with your own prompt, or feed it to ChatGPT 57 times to extract every morsel of meaning.

If you’re not a big workflow person, don’t worry. I have some workflow snapshots that you can use to get up and running.

Workflow Trigger

Example Transcript

As you can see below, the native transcripts deliver a highly reliable breakdown of who was speaking, and at what time. This makes the transcript far more useful when it comes to AI (or even human) analysis.

Speaker Talk Times

The speaker talk time percentages should always add up to 100 (as in 100%).

This is a new feature with this app using the native transcripts, since it wouldn’t be possible with the limitations of a third-party transcription.

I included it because I think it’s useful for sales people or sales managers to be able to see how long each person is speaking for.

If you’re not aware, as a salesperson your goal should not be to dominate the conversation. You at least want to be below 50% of the total speaking time.

Example Workflows

Workflow Snapshots

There are two ways to install this app: at the agency level, or at the sub-account level. As such, there are two different apps.

These snapshot work exactly the same, but they have different triggers, and thus different snapshots.

When you install this snapshot, be careful to only import the workflows!

There is a dashboard, and a few custom (default) fields that can’t be removed. If you import these extra components it may mess up your sub-account configuration.

If you have any touble with importing or setting up your snapshot, feel free to let me know or book a call.

Just note that without the proper permissions, it may not let you import the sub-account snapshot.

Post-Installation Pages

Agency Installation

👋Thank you for installing!

Sub-Account Installation

👋Thank you for installing!

Frequent Questions