Atlassian on Thursday introduced that it has acquired Rewatch, an AI-powered assembly notetaker, display screen recording device and video hub. The firm plans to combine Rewatch into Loom, the asynchronous video messaging platform it acquired for $975 million final October. But perhaps much more importantly, it additionally plans to combine it into its not too long ago launched Rovo AI platform in order that assembly notes can rapidly develop into Jira points, for instance, and the transcripts develop into searchable inside the general enterprise context.
The two corporations didn’t disclose the value of the acquisition. Back within the heady days of early 2021, when all people was seemingly in search of higher video-centric options, Rewatch raised a $20 million Series A spherical led by Andreessen Horowitz. The firm didn’t elevate any further funding since, in keeping with Crunchbase.
As Loom co-founder Joe Thomas instructed me in an unique interview forward of Thursday’s announcement, Loom already provided an integration with Zoom to report conferences and create transcripts. But that solely felt tangential to what the corporate was making an attempt to do with Loom, he mentioned.
“It was worth evolving our platform and bringing in great IP from Rewatch to accelerate our efforts there,” Thomas unhappy. “The reason why we got so excited about Rewatch is the fact that we believe Atlassian is disproportionately positioned to take meeting recordings and maximize the value of them. This is because Loom already generates transcripts for every single video that’s created and then we layer all the AI prompts around it — part of Rovo is that it is a unified search [platform] across an enterprise and it’s also building on top of it.”
Loom and Rewatch share a typical investor in Andreessen Horowitz; Thomas and Rewatch founder Connor Sears sometimes discovered themselves in the identical conferences a couple of years in the past. But it was Atlassian’s company growth group that first met with Rewatch after which requested Thomas to fulfill up with the corporate, too.
Thomas believes that integrating Rewatch’s tech stack will truly be fairly easy, particularly now that Loom has moved its stack over to the Atlassian platform.
The actual problem now could be to offer one of the best person expertise, he mentioned. Once an agent joins a gathering, it has to grasp a whole lot of context; regardless that this shall be a human-in-the-loop system, it nonetheless has to get it proper more often than not. If it continuously suggests the fallacious motion gadgets after a gathering, for instance, customers will rapidly quit on it.
“Rovo and agents is, I think, relatively technically complex but also end-user complex in terms of, OK, if we are talking about a Confluence Doc, what parts of that am I updating? What is actually valuable for an end user or meeting attendees to do on their behalf, from an AI agent perspective? That is something that is relatively new for a lot of us at Atlassian. … I think that that’ll probably take a good six to 12 months to really lock in on that front, with a lot of experiments between now and then,” Thomas mentioned.
One different Rewatch function that Atlassian was particularly concerned with is its calendar integration. The Rewatch group made {that a} very easy expertise, the place customers can merely toggle the assembly bot on and off for every assembly. Rewatch additionally options quite a lot of automation options that, for instance, ship out the assembly notes to each attendee. Indeed, Thomas famous that whereas the Rewatch group confronted a whole lot of challenges in constructing the product, constructing the calendar integration meant coping with extra edge instances and extra friction than constructing most different components of the Rewatch stack.
Once the mixing is full, the Loom AI agent will be capable of be a part of Zoom, Google Meet and Microsoft Teams conferences. It’ll create a full transcript, assembly notes and motion gadgets that it could possibly then robotically correlate to Confluence pages, Jira points and…
