There’s no question that we’re living in a golden age of audio transcription.
Even as I write this article, I’m not even typing. I’m dictating on my laptop.
Siri, ever dutiful, transcribes my words and punctuation just as quickly as I speak them, and with few errors.
Why am I transcribing audio instead of typing? I write almost 50% faster by voice than by finger.
That’s not a guess. I’ve charted my typing speed versus dictation by tracking word counts over 50+ writing sessions. Despite typing 90+ words per minute, voice typing is still faster.
That kind of speed and efficiency are exactly why many professionals turn to meeting transcription software and apps to do their note-taking in meetings. Instead of taking meeting notes by hand, transcription offers a handy (and hands-free) alternative.
It’s a seductive value proposition: focus on the conversation, on the people at the meeting. Let the AI take meeting notes for you.
But how does meeting transcription work in reality? Is it really better than taking meeting notes by hand?
Speech-to-text technology for audio dictation has come a long way, but meeting and call transcription is a wholly more complex problem than dictating a blog post.
Right now, I’m in a quiet office, speaking clearly, about eight inches from a quality Blue Yeti microphone to record the words you’re reading. I am the only one speaking, and I never interrupt myself or make any background noise.
Can meeting transcription software handle poor quality audio, with multiple speakers who sometimes interrupt each other? What about the background noise: the sirens, the door closes, the chair scoots, and all the other extra noise you get in a meeting?
At best, inaccurate transcription yields frustrating meeting notes that are hard to skim or understand. At worst, bad transcription software is worse than useless, putting out thousands of words of transcription that will never be put to use and only clog up your work tools.
For the research behind this article, I put the top four meeting transcription apps to the test, comparing features, pricing, and the quality of the transcription.
Based on rankings and reviews on G2, Capterra, and in the media, the top meeting transcription platforms are:
Some of these apps do more than simply transcribe meeting audio, which is why they call themselves things like “Conversation intelligence platform” and tout benefits like Sales Coaching.
This is because transcribing calls isn’t a challenge that exists in a vacuum. A transcription isn’t a nicely-formatted set of notes taken by a human scribe. It’s everything that has been said.
Have you ever read the transcript from a big trial?
Even for the most famous, most interesting court cases, the transcriptions aren’t nearly as compelling as what actually transpired.
If you want to read an exciting legal thriller, don’t go read court transcripts. You’ll be bored to tears. Go find a book by John Grisham or Chris Pavone.
Meeting transcripts are somewhat similar, in that they don’t replace meeting notes across the board. Transcripts are clunky and often boring, and are only useful if you’re willing to skim past all the unnecessary information to find the meat.
In fact, most of the time, instead of a transcript, you’re better off jotting down 10 bullet points from what happened in a meeting. Otherwise you end up with way too much information—so much that it’s simply not useful.
There are, though, a handful of really good reasons to transcribe calls and video conferences that make a lot of sense.
Reasons to transcribe a call or meeting vary fairly widely depending, especially based on what role you’re in at work. Here are the most common reasons that people record and transcribe calls.
As you compare tools across the meeting transcription market, you’ll see a common theme. These apps are used by sales people more than any other role.
Sales people are busy, on the go, and are notoriously unwilling to take quality notes during sales meetings.
Not wanting to take notes is for a good reason—in sales, your relationship with the prospect is everything. You need to be an excellent conversationalist and an excellent listener. Taking meeting notes splits your focus and can even be construed as rude.
Benefits for transcribing sales calls:
Similar to sales people, customer success reps have to juggle a lot of moving parts during customer calls. The benefits for account managers and customer success reps are similar to sales.
Product managers who are doing customer research also often benefit from transcribing these calls. Figuring out the nuances of a customer's challenge is a detailed endeavor, and having a person's exact words in your meeting notes can come in handy.
Reasons to transcribe customer calls:
In recruiting, the use case for meeting transcription software is two-fold. One is that recruiters tend to have a lot of meetings. They’re chatting with people all the time, and often having similar conversations over and over as they try to fill competitive roles.
Recruiters, like salespeople, also need to pay keen attention during their meetings. Since recruiters already have a bad reputation as not caring about their candidates (only the commissions they’ll get if they fill the role), a good recruiter will pay extra attention to giving a positive impression.
Why recruiters transcribe interviews:
Most internal meetings don’t require transcription. Well-organized teams show up to internal meetings with an agenda, and have at least one person, if not multiple, helping to take notes. To transcribe the meeting would simply be way too much information.
However, for executives, or in organizations with strict legal requirements, transcription may help solve specific challenges around documenting important meetings.
Reasons to transcribe internal meetings:
There are a handful of other reasons to transcribe calls and recordings. We won't spend too much time on them in this article, but if you fall into one of these groups you may still find the information here useful:
Now that we understand how transcription is used in most businesses, let’s compare the various meeting transcription software alternatives and see how they rank up against each other.
The top meeting apps fall into two pairs.
The first question is: Do you want to record, transcribe, and analyze your sales calls with a beefy, enterprise-grade solution?
In that case, compare Gong.ai vs. Chorus.ai.
Or do you simply want to transcribe meetings for any purpose, sales or otherwise? Otter.ai vs. Fireflies.ai is probably more suitable to your needs.
Gong is a sales-focused meeting transcription tool that describes itself as a Revenue Intelligence platform.
Gong’s AI analyzes the text from interactions across video calls, phone calls, emails, and even messages, applying advanced business intelligence algorithms to spot patterns and help sales teams win more deals.
Instead of relying on opinions about what is working in your sales conversations, Gong promises analytics based on data (not guesswork).
For example, if top performers talk more about “security” and less about “discounts”, Gong will make a recommendation if a salesperson strays from this template.
Gong may transcribe meetings for you, but that’s not the value that is referenced in most of its reviews. Fans rave about the “insights” and “tips and tricks”.
This tool faces the reality of meeting transcription head on — that nobody really wants to look at the transcripts. Even if the overall transcript has mistakes, it doesn’t matter so much. Instead, Gong wants to use AI to help you find patterns in the data.
Similar to Gong, Chorus is heavily focused on sales applications with a Conversation Intelligence platform that helps quantify trends like feature requests and conversation topics.
Almost all of the same features that Gong offers are also in Chorus. In fact, reviewing both tools, it’s extremely hard to tell them apart.
At the time of writing, both transcription apps help sales teams in a very similar way, although Gong’s offering appears slightly more robust than what Chorus has to offer.
For either tool to work for your team, you’ll need to be a large organization—Gong and Chorus need data to make suggestions. Your ACV (average contract value) will also need to be high enough to justify the costs of spending thousands per user per year.
If spending 5-figures per year on meeting transcription costs is outside of your investment range, there’s good news. There are two transcription alternatives to compare that come in at a much lower price point, including free.
Overall, both Gong and Chorus are extremely sophisticated in their product vision, and hyper-focused on a single part of the market: sales. Users cite imperfections in the transcripts created from both tools, but you don’t need to look at any of your meeting transcripts to get valuable insights.
Gong is the premium choice here: A little more expensive, and a little more powerful feature set than Chorus.
On the other end of the spectrum from Gong and Chorus (the loud-sounding words for company names) we have Otter and Fireflies (two cute animals).
Otter and Fireflies are more laser focused on a specific problem: meeting and call transcription.
You might consider Otter and Fireflies simply AI-powered meeting notes or meeting minutes transcription services, although they layer on some modest additional features such as tracking keywords and topics.
At any rate, if you need a text version of what happens in your meetings, Otter and Fireflies may have you covered.
Otter has been quickly making a name for itself as a fairly accurate and affordable transcription option for meetings and calls.
Said one recent reviewer on G2: “I was amazed at how well it worked even with people with accents: Boston accents, southern drawls (even Texas style), New York accents, Minnesota -you bet, even handles California valley speak). Everything else is meaningless if it can't do its primary function which is to transcribe speech.”
Still, like all tools in this category, users describe it as “accurate, for the most part” and “The voice recognition is a B-”. Otter is pretty good, but it’s not magic.
Not just for meetings, Otter is sometimes used by writers to dictate out text. This is especially popular with Otter’s mobile app, for taking on-the-go notes.
Otter is easy to try. The free plan will help you understand if Otter is for you, although its many limitations make it challenging to use in the long run.
Otter works with the dominant video conferencing solutions Zoom and Google Meet, but also connects to other tools like Microsoft Teams and WebEx.
At the time of writing, Otter.ai is rated 4.8/5 stars from 22 reviews on G2.
Fireflies is a conversation tracking platform that can automatically record, transcribe, and analyze calls and meetings. It’s a relative newcomer compared to other tools on this list, but has been getting a lot of attention recently.
Fireflies, being newer to the market, lacks reviews on popular sites like G2, which makes it difficult to ascertain how satisfied customers are, or to get broad opinions on its transcription accuracy.
Unlike Otter, Fireflies doesn’t integrate with Microsoft Teams or WebEx, although more and more teams use Zoom or Google Meet anyhow. However, with its Slack, CRM, and Zapier integrations, Fireflies makes it easier to get meeting transcript documents into the app where they really belong.
Fireflies free plan is a little bit less constrained than Otter's. It should give you a good sense of whether Fireflies is right for you.
Otter and Fireflies are likely to duke it out for market dominance in the meeting transcription world over the next few years. While Otter has a head start, Fireflies is hot on its heels. Both tools are similar in functionality and price.
Transcription sounds amazing, but the reality is that most of these tools aren’t accurate enough to be as helpful as you might think.
Every single meeting transcription has a multitude of errors, many of which render a sentence impossible to understand without returning to the audio. What is supposed to be a quick way to skim what happened in a meeting can just as easily turn into a mystery as to what was really said.
The gap between expectations and reality may be explained by the eagerness to apply AI to so many business problems when the technology hasn’t matured enough.
Looking at transcription through the lens of the Gartner Hype Cycle, AI-powered meeting transcription appears to be somewhere between the Peak of Inflated Expectations and the Trough of Disillusionment.
The relatively weak call transcription accuracy is also due to how many complicating factors you have in meetings and calls:
It’s also due to the fact that AI doesn’t hear sound like your brain does. It’s not like being in a meeting as a human. Computers can’t always tell what part of the sound waves are speech and what are something else. The AI doesn’t see who is talking, know where the sound is coming from, or have a true understanding of the meaning of specific words.
You might consider meeting transcription in a similar space to self-driving cars in our current era. The technology sounds awesome and works under ideal conditions, but the world is far more complex.
Oftentimes the best solution for meeting transcription is actually not transcribing at all.
If you’re not in Sales, Customer Success, or Recruiting—and even if you are—it’s likely that your meeting transcripts will be more information than you need or can use.
Say you have 10 hours of meetings a week. Transcribing them all would amount to 78,000 words of transcription each week, the equivalent of a hefty paperback novel in total text.
To retain insights and data from your meetings, it’s not this kind of high quantity of information that you need in your meeting notes. You hardly need any of the contents of the transcription at all.
For most meetings, people need to retain only three things:
Taking notes on these three is quite simple.
What was discussed in the meeting? If your meeting has an agenda, you automatically begin with a list of what will be discussed. Simply copy the agenda as the starting point for your notes and you’ve already recorded a third of the needed information.
If you don’t have an agenda, or your agenda is light on the details, try to build a culture of creating better agendas. Here’s my favorite tip that will make your meeting agendas twice as good.
What was decided in the meeting? Recording decisions is quite simple as well. You don’t need to not verbatim everyone’s thoughts and key points, just their outcome.
If you take notes using the Vital Meetings framework, highlight each decision that is made so that it stands out.
Who is responsible? It takes almost no time at all to put someone’s name down next to a task or action item. If an action is worth taking as the result of a meeting, it’s worth assigning to someone and making a quick note of it.
Just as there are meeting transcription apps to help you transcribe your meetings, there are note-taking apps that are built for meetings as well.
Taking notes using Google Docs, or Evernote, or any other standard note-taking app has its pitfalls. The notes are easily disorganized, hard to share, and often — just like meeting transcripts — after a meeting, the notes are essentially never looked at again.
To avoid this problem, we recommend using a tool like Hugo. Hugo is a collaborative meeting management tool for teams.
Why take your meeting notes in Hugo:
Considering that Zoom itself offers transcription alongside its video conferencing, why isn’t it included in this roundup?
It’s true, Zoom’s recording transcripts feature is available to users starting at the Business tier plan ($199/year/license). This feature automatically transcribes your meeting recording and attaches the transcript as a file in a list of recorded sessions.
In a Tech Target article, news writer Johnathan Dame quotes Zoom as targeting 89% transcript accuracy under ideal conditions.
With its middle-of-the-road accuracy, and lack of AI to help with topic tracking or keyword analytics, Zoom’s offering appears less favorable versus more dedicated tools.
Yes, if you’re already a Zoom Business or Zoom Enterprise subscriber, you might want to leverage a feature you already have, but if you’re thinking about doing a lot of transcribing, it’s probably better to let Zoom keep at what they’re best at — video and audio quality — and use a dedicated tool on top of that for more accurate transcription.
But what should you do if you only need to transcribe a single meeting? Or just a few?
While not specifically for meeting transcription, Descript is worth a mention in this article, and with 3 hours of transcription included in its free plan, it might just solve your problem in a pickle.
Descript is an audio/video editor that includes transcription, screen recording, multi-track editing, and some incredibly useful AI tools. It’s primary users are Podcasters, so you’ll find some features that you may not need, but if you can work around them, give Descript some consideration.
Descript claims industry-leading transcription accuracy, near-instant turnaround, and costs just pennies per minute. Having used Descript on many of my own projects, I can attest that, at least for me and my setup, I find their AI very quick and accurate when it comes to speech-to-text.
And, while nobody is going to host a meeting in Descript, if you have a call recording that you need to transcribe after the fact, Descript (and it’s free plan) may be an excellent option. Import your video, let Descript do it’s thing, and voila! It’s easy to play back and follow along, fixing any errors in the transcript if you need 100% accuracy.
In fact, Descript offers both AI and human-powered transcription, depending on the level of accuracy you need. The robots are pretty good, but the humans are extremely accurate.
It can detect multiple speakers automatically, and you can label their names in your transcript.
If you want to edit your documents at all, Descript gets really interesting. By editing the text in Descript, and you edit the audio, too. Delete a word from your transcription, and it’s gone from your recording.
Using Descript like this, you can automatically remove pauses and filler words like “umm.” You can trim off small-talk from the beginning of your meeting, or even remove entire sections of the meeting easily just by skimming through the words.
With the prevalence of virtual assistants like Siri, Alexa, and Cortana on so many devices, it would be convenient to use their speech-to-text features to transcribe meetings too.
Unfortunately, this isn’t going to work.
These tools are designed to transcribe much short bursts of speech, less than a minute at a time. When I dictate using Siri, I can easily watch the words on the page, stop and start when convenient, and fix mistakes as they occur.
Siri, Alexa, Google Assistant, and Cortana don’t transcribe accurately when there are multiple speakers or blips in audio quality. Plus, your phone is just as likely to go black and stop recording during the meeting because you’re not actively using it.
One of the biggest complaints about meeting transcription is that the transcribed documents are too long to be helpful.
Information overload leads to these docs being created but never used.
A typical speaking pace is about 130 words per minute. Assuming a meeting is no different, the math works out as follows:
3900 words is only for a half-hour meeting. Go all the way to an hour, and you’re up to 7800 words.
7800 words sounds like a lot, but just how much text are we talking about in this transcript?
To put it in perspective, a one-hour meeting transcription, printed out, would amount to about 30 pages, assuming a standard-sized letter paper, double-spaced, using 12 pt font and 1” margins.
At 7800 words, your hour-long meeting’s transcript takes about 26 minutes for the average reader to read through. It’s still a huge time investment.
In fact, when we tested this rule of thumb, we actually found that all of our transcripts were longer. See the chart below. One hypothesis is that people speak more quickly in meetings than under other circumstances in an effort to be more efficient with each other’s time.
So yes, while reviewing meeting transcripts versus attending the meeting may give you a net gain in efficiency, it’s often more trouble than it's worth.
Put another way, this article, which goes into considerable depth, may seem extremely long compared to what you’re used to seeing on the web. You’ve probably skimmed a lot of it to make it this far. The entire thing is a monster at 4,400 words.
Now, imagine reading this web page, top to bottom, two times for EVERY meeting transcription.
Across all meeting transcription software—Gong, Chorus, Otter, and Fireflies—accuracy is a continual problem. Gong and Chorus try to solve this issue by not focusing so much on the entire transcription, but rather by getting value from keywords and trends found in messages.
And even when meeting transcriptions are accurate, they are still a massive amount of text. In most cases, you're better off jotting down 10 bullet points than having 10,000 words you'll never revisit.
At the end of the day, if you’re not using the transcription for a specific reason, like you’re a journalist or product manager who needs the actual words from the meeting, you’re most likely better off taking notes by hand for your meetings using meeting management software like Hugo.
Meeting notes don’t need to be a verbatim accounting of what happened: just include what was discussed, what was decided, and who is responsible for next steps.
Evernote is popular and full of features, but is it really the best for all people, in all circumstances?
Meeting agenda templates to copy or download (Google Doc or Word Doc) — plus examples of how to use them.