Slack Reimagined - video transcript Ken Johnson | LCA interview product-strategy concept Transcribed from the final 6:06 video. Scenarios and savings are illustrative, not measured results. We will make most conversations unnecessary. Slack's mission is to make work life simpler, more pleasant, and more productive. Our original promise was to kill email, but we ended up creating a different type of problem. Endless channel noise, ping culture, and constant back and forth threads that talk about work instead of doing it. Slack was built on three assumptions about the person using it, but a lot has changed. The first assumption was that people are doing the work. Today, handing off work to AI is normal. Agents can finish a task on their own inside the tools your team already use. The second assumption was your team is online when you are. Now work happens when you're offline. Agents run 24/7. Team mates are globally distributed and keep different hours. And what's waiting for you when you come back is a pile of messages and channel notifications that you need to sort, sift, and orientate yourself around before you can decide if your response is needed. The third assumption was that you belong to one team with one job. Now people are working beyond their usual roles. A designer can analyze customer data. An engineer can draft launch copy. A marketer can build a prototype. So what if Slack didn't just help us talk about work? What if Slack could take a conversation, figure out what needs to happen next, and get it started? There are three elements to make this work. A twin that represents you, a producer that turns discussion into work, and Reins, where you decide what an AI can do on its own and what needs your approval. Let's see what that looks like for Janice. She's a product lead at Parcelo, helping customers find contact information for their sales leads. People across the company turn to her for answers, updates, and decisions. She knows the details and remembers why decisions were made, but keeping everyone moving takes time away from her high leverage work. But Janice has a secret. She has a twin. So let's see what's waiting for her on Monday morning. So it's Monday morning, and her twin has already handled 30 questions while she was away. But one still needs her judgment. Janice is launching a feature that automatically finds contact information for sales leads. Over the weekend, a pilot customer uploaded 5,000 leads, but only had enough credits for 2,000. Engineering is asking how to handle this use case. Her twin brings her the options and what they mean for Thursday's launch. Since Thursday is only three days away, she wants to have a call with Jonas to work through the details of the customer experience. The first topic they work through is design. The design agent drafts mockups as they talk, while the producer turns their conversation into Jira tickets, test cases, and support drafts. Before they settle on the solution, they both want to understand what Maya heard from the pilot customer. She isn't on the call, so the producer asks Maya's twin to bring the feedback they need to make the right decision. Janice and Jonas focus on getting the experience right. The agents handle the work around them. Janice approves the solution, and the work prepared during the call is saved and sent to the system of record. She doesn't leave the conversation with the list of follow-up tasks. She just confirms the decision, and the work is already prepared. On Saturday, Devin and Sales ask whether Thursday's launch was still on track. Her twin answered with what it knew, while making it clear that Janice still had a decision to make. Now that she's confirmed, it goes back to Devin with the update, without Janice having to remember who was waiting. But why would Janice let something speak for her? Reins is how Janice stays in control. She can see what her twin has done, correct it, and decide what it can handle without her. As she builds confidence in its work, she can give it more responsibility. Some decisions still need her approval, and company rules still apply. But she doesn't have to trust it with everything to get the most value from it. People are getting used to asking AI before asking a colleague. With the twin, they can get answers from Janice's knowledge without interrupting her. We're also moving away from giving AI individual tasks to trusting it with more autonomous work. Janice makes a decision, the agent turns it into specs, tickets, and whatever else the team needs next. Retrieval helps agents find the relevant context, tool use lets them act on it, and agent to agent communication lets them get answers from another twin. Slack supplies the reasoning behind that work, what was decided, why, and what was promised. That means fewer messages asking for answers, fewer calls to check progress, and less chasing people to get things done. When people do need to talk, the agents turn that conversation into work in real time. Imagine a week with 30 fewer interruptions and zero follow-ups to remember. Your team gets what they need, you get the time and the headspace to focus on where you can make the biggest impact. Slack's mission is to make work life simpler, more pleasant, and more productive. That means fewer interruptions to get an answer, fewer follow-ups to make a decision, and less time catching up before you can contribute. People will still come together for the questions that need their judgment, and while they talk, the producer will turn their decisions into work. And afterwards, their twins will update the people who need to know. That's how we make most conversations unnecessary. By taking care of the questions and coordinations that call to them. Imagine starting Monday ready to contribute because catching up is already taken care of.