Best AI Tools for Transcribing Meetings in 2023

Just about everyone’s heard—and likely felt—the sentiment, “This meeting could’ve been an email,” but in spite of that, in-person meetings and check-in conference calls are not going to stop being part of the business world any time soon. Fortunately, recent advancements in artificial intelligence technology are poised to revolutionize how we retain information shared in meetings, as well as how effectively we can use that information once all participants go their separate ways.

A good AI transcription tool should not only copy exactly what’s said aloud, but also be able to identify distinct speakers, extract important action items and assignments, collate information for easy review and retrieval later, and increase accessibility for people with hearing loss or anyone who may have trouble devoting their full attention to any particular meeting. Here are some of the top AI meeting transcription tools that are available in 2023, and some details about Deepgram’s efforts to push the boundaries of AI technology further than before.

One of the biggest benefits of AI transcription is also one of the most obvious: no matter how quickly you can type, computers will always be faster at transcribing than humans, and computers have seen substantial improvement around real-time transcription accuracy in recent years.. What’s especially important when it comes to AI transcription for meetings with multiple participants, though, is “speaker diarization”—in other words, distinguishing between different people within a single conversation, and automatically identifying which people make each distinct statement.

For example, the Diarize feature within Deepgram’s API generates a JSON response that detects and segments each distinct word within a pre-recorded audio file, and then tags each word coming from a particular “speaker”—Speaker:0 for the first unique voice heard in the call, Speaker:1 for the second, and so on. Notably, Deepgram’s API is also capable of transcribing each channel within an audio file separately, and then diarizing speakers within each channel. Werecently partnered with Vonage to demonstrate Deepgram’s automatic speech recognition (ASR) capabilities in action during a live call.

Other key features you should be looking for in a high-performing AI meeting transcription tool includes: Action Item Extraction, automatic generation of meeting summaries and highlights, and support for multiple languages. Put together, a high-quality AI transcription tool should save you money and improve productivity while helping minimize the strain that manual transcription often brings. 

If you’re just starting to explore your options for AI meeting transcription tools, Otter.AI is a great place to begin. This straightforward and user-friendly platform is designed to provide fast and accurate live transcriptions—as well as smart summaries highlighting important points and takeaways—not only for get-togethers in the business world, but also for students and educators participating in lectures, seminars, and faculty meetings.

The basic version of Otter.AI is free and offers up to 300 transcription minutes per month for conversations up to 30 minutes long, as well as three transcriptions of imported audio or video files. The Pro version is $8.33 USD per month, can handle 90-minute conversations, comes with 1200 transcription minutes and 10 imported transcriptions per month, while the Business version will run you $20 USD per month and provides 6000 monthly transcription minutes, maximum four-hour conversations, and unlimited import transcriptions.

Rev.com is a unique option for AI transcription meetings because, technically, it’s not just an AI transcription service. While this platform does provide automated transcription through its speech recognition AI for $0.25 a minute, it also provides manual text transcription and English captioning for $1.50 a minute, as well as caption translation for 17 other languages with pricing ranging anywhere from $5 to $12 a minute. You certainly wouldn’t be alone if you choose this option either, as Rev boasts over 750,000 users that include major organizations like Spotify, Viacom, and various universities.

Sonix advertises itself as being capable of transcribing 30-minute audio files in as little as three to four minutes, and of doing so in nearly 40 different languages. Between that and the numerous other features provided by this product, including automatic word timestamping and speaker diarization, exporting options for both text and subtitles, and the ability to create custom dictionaries for industry-specific terms, it’s no surprise that Sonix is the AI transcription tool of choice for companies like IBM and Adobe. 

The standard version of this tool has a pay-as-you-go price point of $10 per hour. A premium subscription shaves that rate down to $5 an hour alongside a $22 monthly fee per user, and enterprise pricing is also available for organizations with high-volume transcription needs.

On top of essential AI transcription features like automatic meeting summary generation and searchable voice transcripts, Fireflies also allows users to add tags and comments to transcripts, streamlining collaboration with other team members. Even better, Fireflies can be added to Google Chrome as an extension, and be set up to automatically generate tasks in programs like Asana or Trello.Fireflies can even track important metrics within transcripts like which person is speaking the most. 

Fireflies offers limited transcription credits through its free version, but most of the unique features are reserved for users with pro or business plans, respectively priced at $10 and $19 per seat/month if billed annually. There is also an enterprise option—exclusively billed annually—for teams with 51 or more members.

AssemblyAI’s main selling point is Conformer-1, its in-house speech recognition model that was trained on more than 650,000 hours of supervised audio data. It also boasts enterprise-grade security, 99.99% uptime APIs capable of handling millions of daily requests, and key features like customizable vocabularies and labels for unique speakers. 

AssemblyAI prices its service by the second, offering core transcription services at $0.00025 per second, audio intelligence features like summarization and sentiment analysis for an additional $0.000583 per second, and its LeMUr product—tailor-made for transcribing and analyzing spoken conversations—for an additional $0.000036 or $0.000118 on top of the core service pricing, depending on whether you choose the basic model or the more advanced one.

It probably comes as no surprise to hear that Amazon Web Services (AWS) has also dipped its proverbial toes into AI meeting transcription services. Amazon Transcribe is capable of both recorded audio and real-time transcription and has models fine-tuned for specific types of transcriptions—for instance, low-fidelity phone calls received by call centers. You can also expect to get all the customization, diarization, security, and analytical features one could possibly need from an AI transcription service, all at per-minute prices adjusted for standard, PII redaction, and CLM services.

Speechmatics’ Ursa model is billed as the most accurate speech-to-text system available on the market, as well as one capable of recognizing and transcribing various accents that other models struggle with. As if that weren’t enough, Speechmatic also provides real-time translation services for almost 70 language pairs and can be deployed either as part of your own internal environment or through the cloud. Batch transcription services are priced at $0.80 per hour for the standard model and $1.04 per hour for enhanced features, and real-time transcription is available at $1.04 or $1.35 per hour depending on the model chosen.

Of course, if you need fast, accurate, and cost-effective AI meeting transcription services tailored to an industry-leading standard, need look no further than Deepgram’s AI-powered speech-to-text and speech understanding API. Deepgram is an API-first developer platform provider creating composable building blocks for Language AI. The goal is to give any developer access to world-class end-to-end deep learning (E2EDL) through just an API call.

Deepgram powers transcription for leading companies in the online meetings and meeting-augmentation sector, and developers can try out Deepgram by signing up for an account, which grants access to the API Playground, our LLM test bench, and more.

While every meeting (at least ostensibly) has a purpose, not every meeting is easy to keep your full attention on while you’re occupied with other tasks, or simply trying to parse a great deal of information all at once. AI transcription tools can make problems like this a thing of the past—and on top of that, make your business more efficient and more profitable as a result.

Deepgram has the top AI meeting transcription services you may need to keep up with the pace of the modern business world. Get in touch with us today to discuss your options.