How to use AI to automatically summarize meeting transcripts

The use of virtual meetings by businesses increased from 48% to 77% from 2000 to 2022. While initially spurred by the pandemic, virtual meetings have now become a critical component of business communication.

Internally, virtual meetings are used by a continued remote workforce, but externally, virtual meeting platforms have popped up to help facilitate virtual sales calls, virtual interviews, and additional forms of virtual communication.

These virtual meeting platforms are often comprised of impressive AI-powered tools and features that augment communication, collaboration, and sales engagement by automatically summarizing meetings, extracting key insights from meeting notes, or recommending next steps post-sales calls or interviews.

In this article, we will examine one component of virtual meeting platforms: using AI to automatically summarize meeting transcripts. We will look more closely at what AI summarization is, how it works, and the top benefits of AI summarization. Finally, we will discuss how to automatically summarize meeting transcripts yourself, or how to choose one of the top companies to use to automatically summarize meeting transcripts for you.

What is AI summarization?

AI summarization models automatically distill text or transcripts, like from a document, sales call, podcast, or research paper, into their most important parts.

AI summarization models can be purpose-built for various purposes, such as the AssemblyAI Conversational Summarization model that summarizes conversational texts or transcripts. This model is trained specifically on conversational data, making it more accurate and useful for summarizing interviews, customer/agent calls, and more.

How does AI summarization work?

Text summarization methods are categorized into two groups: Extractive and Abstractive.

Extractive summarization models extract what they deem the most important sentence/s directly out of the original text. This method does not alter the original language used in the text, instead compiling the summary from the original language of the text.

Abstractive summarization models use AI to generate original summaries from the text. This method produces summaries that capture the salient information in the text, which may or may not include words and sentences from the original text.

This blog post explores how these different summarization methods work in further detail.

What are the benefits of AI summarization?

AI summarization has many benefits for virtual meetings and other use cases.

For example, AI summarization is an important addition for many Conversation Intelligence suites as it can help facilitate:

Faster QA and call review
Identifying key trends among aggregate call data
Monitoring calls for key insights
Accurately summarizing calls for record-keeping
Increasing representative engagement by minimizing note-taking

And more.

AI summarization for call coaching shares many of these same benefits and can also help enable better context sharing between product, marketing, and other teams that are invested in customer data.

How to automatically summarize meeting transcripts with AI

If you would like to automatically summarize meeting transcripts using AI yourself, there are several approaches you can take.

Here are a few tutorials to follow, though note that some require knowledge of coding to implement:

Summarize audio transcripts in 30 seconds with Python

This video tutorial teaches you how to obtain a free AssemblyAI key, transcribe an audio or video transcript, and obtain a summary of the transcript using Python. You can also change the type of summary generated and the summary model used by adjusting the specified parameters.

Get a custom LLM summary of your audio files with LeMUR

This video tutorial demonstrates how to use LeMUR, AssemblyAI’s framework to process audio files with a Large Language Model (LLM). The six-minute video teaches you how to effectively use LeMUR’s custom summary endpoint to get summaries in any way you would like, including TL;DR, bullet points, and short sentences.

Build a Streamlit app to summarize podcast episodes

In this video tutorial, you learn how to build a Streamlit app that automatically summarizes podcast episodes (or recorded meetings).

How to summarize meeting transcripts with no-code AI

There are also a few additional ways to summarize meeting transcriptions using no-code AI applications.

Use ChatGPT for meeting notes

If you’re interested in summarizing meeting transcripts without any coding, one approach is to utilize ChatGPT. In order to get the meeting transcript, you’ll first need to have a transcription of your meeting audio or video file. This can be obtained by running the file through a Speech-to-Text API such as AssemblyAI.

Then, you’ll need to prompt ChatGPT to provide a summary of the transcript with a prompt such as “summarize the key points of this meeting.” The more specific the prompt (i.e., provide the summary in bullet points) and the more context you provide (i.e., this is a performance review meeting), the better the summary will be.

A more detailed tutorial on using ChatGPT to provide meeting notes and summaries can be found here.

Upload your meeting audio file to AssemblyAI’s playground

Alternatively, you can obtain a detailed summary without the extra step of transcribing the file first using a no-code AI playground like AssemblyAI.

Simply upload the audio file you would like then toggle on AI models and/or endpoints to output a summary that best matches your requirements. Options include:

Summarization: Summarize your audio file into bullet points, a gist, headline, or paragraph using one of AssemblyAI’s summarization models (informative, conversation, or catchy).
LeMUR: Leverage LLM capabilities to create custom summaries of your audio data.
Auto Chapters: Summarize your audio file into chapters with timestamps.

Then, you’ll receive both an accurate transcription of your meeting, as well as a summary based on your chosen model and/or endpoint above.

Try AssemblyAI’s playground

Top AI companies to use to summarize meeting transcripts

If you’re looking to summarize virtual meetings in bulk, or would like to take advantage of additional AI-powered tools and features on top of AI summaries, these AI companies offer advanced AI summarization tools to try.

Fireflies.ai

Fireflies.ai is an AI voice assistant that helps users transcribe, summarize, take notes, and complete additional actions during and after virtual meetings.

Its AI assistant integrates with leading virtual meeting providers such as Google Meet, Zoom, Microsoft Teams, and WebEx and integrates with business applications and CRMs such as Hubspot, Salesforce, and Slack.

Fireflies AI Super Summaries tool automatically provides all users with detailed meeting overviews, outlines, meeting notes, keyword lists, and post-call action items.

The summarization feature helps support easier collaboration between teams and allows for other Conversation Intelligence analysis tools to be applied to the summary.

Sembly AI

Sembly AI is an AI team assistant that transcribes, takes meeting notes, and generates intelligent insights for virtual meetings.

Like Fireflies, Sembly AI integrates with Google Meet, Microsoft Teams, and Zoom. It is also available in 35 languages across web, iOS, and Android mobile apps.

After each virtual meeting, Sembly’s meeting notes view presents a focused digest of the meeting, displaying a short summary, meeting outline, key topics, and meeting insights.

Its Semblian tool also lets users ask questions or auto-generate post-meeting tasks like composing an email or next steps.

Grain

Grain is an AI-powered meeting recorder that also integrates with popular online meeting platforms to record, automate note-taking, and capture insights from internal and external conversations.

Grain’s automated meeting notes tool provides a clickable timestamp with a summary and key points for each call. Users can also highlight key moments in a transcript and use AI to summarize what was discussed during those moments. Users can also create shareable video clips from these key moments.

Grain can also send automated meeting notes to tools like Slack, Hubspot, and Salesforce.

CallRail

CallRail is a lead intelligence software company that helps customers build more meaningful relationships on sales calls, increase ROI on sales tracking, and drive overall community impact.

Its AI-powered Conversation Intelligence feature summarizes each call processed through its platform into a few sentences, helping users understand calls at a glance. Additional features include sentiment analysis, identifying industry trends, call scoring, and more.

These AI-powered features, including AI Summarization, helped the company double the number of its Conversation Intelligence customers.

Callbox

Callbox is a global lead generation company focused on helping users grow sales pipelines with both human and AI-powered tools. Callbox uses a multi-channel approach to reach prospects, generate MQLs and SQLs, and convert more leads into closed deals.

Its AI Summarization feature summarizes all calls into short snippets in its platform. Callbox also integrates with Hubspot and utilizes SMART Calling to better reach target prospects at the best times for engagement and conversion.

Jiminny

Jiminny is a leading Conversation Intelligence, sales coaching, and call recording platform. Sales and customer success teams use its platform to manage and analyze all conversational data, helping the teams fine-tune sales techniques, build better customer relationships, and secure 15% higher win rates for its customers.

The Jiminny notetaker joins each call, whether it be on Zoom, Google Meet, Microsoft Teams, or another platform, and records and transcribes the conversation. Then, the Jiminny platform summarizes the conversation into its most important moments, as well as performs other forms of analysis. Users can even use the new “Ask Jiminny” feature to submit questions about the call and receive AI-generated responses.