# Claude Code in Slack changes how teams SHIP ## Метаданные - **Канал:** Brian Casel - **YouTube:** https://www.youtube.com/watch?v=r5Mzl3ulxDc - **Дата:** 09.02.2026 - **Длительность:** 7:36 - **Просмотры:** 6,524 - **Источник:** https://ekstraktznaniy.ru/video/10949 ## Описание I came across an article about how Anthropic uses Claude Code in Slack internally — and it changed how I think about the ROI of adopting AI with your team. Your engineers are shipping faster with Claude Code, but that intelligence is locked in their terminals. This video shows three ways to extend codebase knowledge to product, support, marketing, and sales — so your whole team can move faster, not just your engineers. 👇 **Your Builder Briefing (free)** https://buildermethods.com - Your free, 5-minute read to keep up with the latest tools & workflows for building with AI. 👇 **Build with Claude Code (new course)** https://buildermethods.com/pro - Builder Methods Pro members get access to my 2026 course on building with Claude Code, plus our entire library of video training for builders and our community of pros. 👇 **Use Agent OS** (free open source): https://buildermethods.com/agent-os 👇 **Use Design OS** (free open source): https://buildermethods.com/design-os ▶️ Related videos: ## Транскрипт ### Claude Code Adoption [] Your engineering team has probably already adopted cloud code and you're probably already seeing the results. They're shipping faster. They're solving problems quicker. They're moving at a velocity that didn't even exist 6 months ago. But that transformation is contained. It's living inside the terminal windows of your developers. So, it's benefiting them, but only [clears throat] them. And meanwhile, the rest of your organization is still operating the old way. Product is still scheduling meetings to hash out features. Support is still waiting on engineers to explain how things work. go to market, it still can't get answers without filing a request. All that intelligence about your codebase, your product, it's locked away, accessible only to the people who can open a terminal, which means that your ROI from your AI investment is a fraction of what it could be. So, the other day I came across this article from a member of the Anthropic team about how they're using Claude code in Slack internally at Anthropic. you know, they're the makers of Claude. And it opened my eyes to what's possible when you extend codebased intelligence beyond just the engineering team. And so, three ideas stuck out to me about how they work. So, let's see what we might be able to apply to our own teams. Whether you're running a small studio like I am, or you're running a much larger organization. And if it's your first time on my channel, I'm Brian Castle. I help builders and teams navigate this transition and adopt AI. And so, every Friday, I send my builder briefing. It's a free 5-minute read with my real take on what's actually working in AI first development. You can get yours by going to buildermethods. com. And if you're serious about adopting Claude Code this year, I just launched a new course called Build with Claude Code that's available to builder methods pro ### Prototyping new features via Slack [1:37] members. So here's the first thing that stuck out to me from that article. He said, "I don't really send memos or make mock-ups anymore. I just make Cloud Code prototypes. " So, if we think about how feature development usually starts, maybe it's a meeting, maybe someone writes up a PRD or a spec, maybe there's a Figma mockup, maybe a product manager even vibe codes a quick prototype to show what they're thinking. Problem is, none of that is real. A mockup can't show you how something actually behaves. A PRD or a spec that's open to interpretation. A vibecoded prototype that looks and functions nothing like your actual product. It's built in isolation. and it's disconnected from your codebase, your design system, your architecture. And so eventually it all gets handed to engineering anyway. And then they start from scratch building the real version based on their best interpretation of what everyone wants. And so the anthropic team does something different. When someone has a feature idea, they tag Claude directly in Slack and ask for a first pass prototype. That's an actual pull request built against their production codebase. So that's not a throwaway demo, not an isolated sketch. It's real code using your real components, following your real patterns. Now, your product team might not ship that PR as is, but it's dramatically more productive for them to react to a working version in the actual codebase than to interpret a PRD or a mockup. So, they could pull it down, they could run it, they could see exactly how it behaves, and then if they need to, they could rebuild it properly with that clarity. The fidelity of the conversation here changes completely. So anyone on your team can kick this off, whether they're in product, support, sales, they don't need to wait for engineering bandwidth just to explore whether an idea is worth pursuing. When this works, you're going from idea to productionready feature in a matter of days rather than weeks. But more importantly, you're nailing that exact job to be done that your customers actually need because you're iterating on something real from the start. So just by integrating Claude with Slack, you've turned this into a real win for your customers, your team, and your ### Your team has code questions [3:38] business. All right, here's another idea from that article. When people from the go to market team, the customer support, marketing product, when they have a question about the product and how it works under the hood, they ask Claude directly in Slack. Now, that sounds simple enough, but think about what that's actually replacing. You've probably experienced this, where your support rep is on a call with a frustrated customer. They need to know how a feature handles a specific edge case right now. So they ping an engineer and they wait and then the customer waits and then the tension builds. Or your marketing writer is finishing up a blog post or a video about a recent release. So they need to know when it shipped and what was the actual reasoning behind it. You know what makes it interesting to your customers. So to get that information, they would need to ping an engineer and maybe their deadline slips while they wait for that response. or your sales rep just got off the call with a prospect and they asked something technical. So now they have to track down an engineer, wait for a response, and hope that the answer comes back before the prospect moves on to evaluating a competitor. Meanwhile, the engineer, they're deep in architecting that next feature or refactoring some business logic when another notification pulls them out, another question that they've already answered before, another context switch that breaks their momentum. Claude and Slack changes this. Your team can ask questions directly and get answers based on your actual codebase including git history like how does this feature handle this edge case or when did we ship this and who built it and why does this work the way that it does. The support team can resolve customer issues faster. Marketing is shipping on deadline. Sales can follow up while their conversations are still warm and engineers stay focused. their investment in good code and documentation that's now paying dividends across the organization without them needing to stop to explain it. Right? This one might be the most ### Business insights via Claude for Slack [5:23] underrated idea and it comes from this pro tip he sort of buried in the article. He said, "Go even further and add Claude skills with access to your event store or analytics dashboard to answer questions about feature usage or error locks. " Think about what that unlocks. Your product lead wants to know how many users completed a specific flow last month. And normally they would ask an engineer what event to look for and then ask a data person to run the query and then wait for both of them to have time. With Claude in Slack connected to your codebase and your analytics, they just ask. Cloud knows what events exist because it can read the code and then it can query the data directly. So questions like what's the completion rate on the new onboarding flow and which features are driving the most engagement this month? So your ops lead can connect code changes to business outcomes without guessing. Your CEO can get clarity without scheduling a meeting. So your codebased knowledge and your business data is finally in the same place. And so that means that more of your team is moving the needle instead of waiting in line for answers. So this is where the real leverage is. Not just answering questions about code, but combining what your code knows with what your data shows. So to recap, these ### 3 use-cases for Claude in Slack [6:35] are the three ways to extend cloud code beyond the terminal and into the rest of your organization. Feature ideas that start as working code, technical questions answered without the engineering bottleneck, and business insights that combine what your code knows with what your data shows. The common thread is your codebase becomes a shared resource. Not something locked away in terminals that only engineers can access, but it's actual intelligence that powers your entire team. Now, if your engineering hasn't fully adopted cloud code yet, or if they're still figuring out how it fits into their workflow, I've got a video that might help. You know, because there's a lot of noise out there about frameworks and elaborate setups and all the things that you supposedly need in order to be productive with cloud code. So, in that video, I break down why it's actually simpler than the hype would have you think and why vanilla cloud code is all most teams actually need in 2026. So, right after you hit subscribe on the channel, I'll see you over there next. Let's keep building.