How Perplexity Computer Turns Slack Into an AI Teammate (Full Transcript)
[00:00:05] Speaker 1: My pitch about computer is being able to bring all the tools that you already work with into one place where you can also collaborate with your team, and that place is Slack. Before computer, things were pretty painful. Every single process had a different workflow, different request process.
[00:00:25] Speaker 2: So much of knowledge work is stitching together many different systems in order to get things done, and now instead of having to go and manually bounce around a dozen different tabs in your browser, you can do all this right in one spot. Slack is where enterprise work is happening today. So as we were building computer, Slack was just the obvious place for us to start. Within a matter of a few days, we had a few hundred lines of code thrown together very quickly to wire up our agent harness to Slack APIs. We took computer from initial concepts to launching in six weeks. One of the things that we've been really excited about with computer is plugging into all of your business systems. We have 400 plus connectors, and being able to then, in Slack, at mention computer and just ask it to get something done across whatever workflows you may have. Computer can do all of that and then come back to you right in Slack with a result that otherwise would have had you popping out to a dozen different tools.
[00:01:19] Speaker 1: A really big aha moment was when we were able to connect Perplexity Computer to Salesforce, to Slack, our UXR calls, and be able to paint an entire picture across all these different sources of what the customer journey actually was.
[00:01:34] Speaker 2: Slack is inherently multiplayer, and taking the workflows that are already happening amongst folks on the team and bringing computer right in alongside of that work and having it act as a teammate is really, really powerful.
[00:01:45] Speaker 1: It's almost as if you're adding another team member or even 10 team members into Slack to have it all be multiplayer. For example, on a landing page, if I wanted to change a bit of the copy, I no longer had to bug an engineer. Computer can spin up a PR because it's connected to our GitHub already, and then the engineer can come in and approve it. You can also call on a model council where you can have three different models to poke holes in your argument. Computer can give you a starting point with all the information that is accurate and contextual to you and your team.
[00:02:24] Speaker 2: What really gets me excited over the next couple of years with computer and Slack is unlocking the potential of humans by enabling them to focus on differentiated, high-value work and delegating the rest to computers. I can't imagine computer without Slack. It's how we run every part of the business top to bottom.
- Perplexity Computer uses Slack as a central interface for executing work across business systems.
- The product supports more than 400 connectors, reducing the need to switch among multiple tools.
- The Slack integration was rapidly prototyped and the product launched within six weeks.
- Computer can synthesize context from sources such as Salesforce, Slack, and UX research to map customer journeys.
- It can assist with practical workflows, including opening GitHub pull requests and coordinating model-based critiques.
- The vision is to make AI agents collaborative teammates that handle routine work while humans focus on differentiated tasks.