What’s worse than missing a meeting? Attending one. So say most workers and managers, even as more meetings were held in the last 15 months than perhaps in any previous 450-day period.
Otter.ai wants to free you from taking notes at virtual and hybrid meetings—and maybe even from attending them. Its deep-learning voice-transcription service can now scan your Google or Outlook calendar for Zoom sessions, automatically sign in at the appropriate time, and produce a live transcription that you and other participants can correct, annotate, and highlight in real time. Or—if you’ve ditched the meeting—your colleagues can while you’re busy elsewhere.
Launching today, Otter Assistant automates part of the company’s existing integration with Zoom videoconferencing for real-time captioning and live transcription. With a Business tier plan ($30 per month or $240 per year for each user), Otter Assistant doesn’t need you to host a session or even log in to take an audio stream and perform speech-to-text transcription. You can select which meetings it attends.
“Once you synchronize your calendar with Otter, at the beginning of the meeting, Otter Assistant will automatically join that meeting, then take notes, and also, with your permission, will share those meeting notes with all the attendees,” says Otter.ai CEO and cofounder Sam Liang. At the most basic level, Liang says, this avoids you having to remember to start recording or perform any other setup in a meeting. But that’s the least of what it offers.
As with Otter’s existing live transcription feature, the person whose account is hosting the process can allow others to highlight text, insert comments, and add images. Transcripts are also post-processed to produce higher-quality final versions that improve accuracy and the identification of distinct speakers. When it’s complete, unique speakers can be labeled, and the system updates them throughout the document. You can edit the transcription on its own or while listening to recorded audio synchronized with the extracted text.
Otter’s accuracy isn’t perfect, but it’s among the best available, which includes competition from trillion-dollar firms. The company develops its own cloud-hosted algorithms, which lets it constantly improve them. It also allows them to spin up capacity on demand, which will be required with more real-time usage that the assistant feature will promote. Since the service’s inception, the company says it has transcribed over 150 million meetings representing over 5 billion minutes of audio.
Major makers of videoconferencing services such as Zoom, Microsoft, Google, and Cisco don’t offer their own in-lieu-of-attendance bots, and the captioning and transcription features in Microsoft Teams, Google Meet, and Webex are relatively limited and lack rich collaboration. Not many third parties have offered robust bots yet—the free, rudimentary Zoombot and the subscription-based Beulr (if you’re a Gen-Xer, say it aloud) appear to be the only contenders. While privately held and independent from Zoom, Otter has a close working relationship with the company, which includes Otter service in some of its tiers and offers an easy connection to Otter features.
Great—even more meetings
CEO Liang paints a picture of the transition underway in businesses from entirely virtual meetings to hybrid ones, with a mix of people at home, at remote work locations, and in shared offices. Even with high levels of vaccination and a substantial drop in cases, most companies expect the majority of full-time office workers won’t be back in their spaces anytime soon; a substantial number of firms will have half or fewer of their employees on-site for the next several months—or forever.
That likely means more meetings rather than fewer, though with less critical involvement across the entire meeting by all participants. A survey Otter commissioned in March from YouGov of 2,000 U.S. and U.K. remote workers found that 35 percent felt people “only attend relevant sections of meetings,” 33 percent wanted shorter meetings, and 25 percent said “meeting notes should always be shared with attendees.”
Otter suggests that automated live transcription of most meetings would improve efficiency and retain more information about business decisions. “When you think about the percentage of knowledge that’s shared verbally in companies—in today’s world, most of that just kind of evaporates, unless individuals are taking their own notes,” says Kurt Apen, Otter’s CMO. “Even in that instance, often the notes aren’t shared, and so you can’t collaborate with them in the same way that you would with a Slack message thread or an email chain.”
Otter’s live transcription turns meeting audio into a resource for participants to follow along with, a catch-up tool for those joining late or asked to join at a particular time, and longer-term institutional memory. Apen says the notes become a repository for groups or the entire company, where personal note-taking is often incomplete and audio recordings are too burdensome to review in any quantity. Simply being able to search on key terms spoken can be a killer feature.
Liang says that the automatic availability of a transcript can compensate for the basic limitations of spoken communications, too. He says people often spool out numbers such as revenue, churn rate, and interest rate, but “your human auditory buffer is quite limited in how many you can remember at the same time.” He adds that he’s sometimes asked to attend simultaneous meetings among the 30 to 40 he has scheduled each week. He uses his company’s product to follow along in two meetings at once.
Universal meeting transcription also helps with a global workforce such as Otter’s, the company says, avoiding the need in many cases to get everyone to coordinate around one or more time zones. And Liang notes, “For nonnative speakers, having the live transcript actually enhances their understanding of the conversation.”
At launch, Otter Assistant works only within Zoom, although the rest of its service offering also works within Google Meet. The firm made no announcement about adding other videoconferencing integrations. All tiers of Otter’s existing service allow recorded audio to be uploaded and transcribed, but that’s not quite the same benefit.
What’s next for the Otter Assistant? Liang was willing to speculate a bit about its future. Processed transcripts already put a summary of popular keywords from the audio at the top, which leads to the obvious notion that a machine-learning company such as Otter could provide additional useful features relating to its understanding of the transcript. “Everybody needs a summarization,” says Liang. “I need it myself; other team members need it. Our users have been constantly asking for some form of summarization.”
When I suggest that the ultimate outcome of what the Otter Assistant is doing would involve robots simply meeting without us, transcribing each other, and reporting back, Liang replies with a laugh, “That will be a little bit longer-term. Someday the Otter Assistant will even talk intelligently.”