Slack Release Notes

From code channels to a smarter search, here's everything new in Slack this month.

Summer’s winding down, but the pace inside Slack isn’t. While people are squeezing in one last hurrah, our teams have been shipping — and this month’s drop is one of the biggest yet.

The theme running through all of it: AI shouldn’t be something you do alone in a tab. Work is better when the people and the agents helping you are in the same room, with the same context, able to see what’s happening and jump in when it matters. That’s the problem this release solves — connecting the AI work already happening across your team to the conversations, tools, and records you already live in.

This month, we delivered on that in a big way. Here’s what’s new.

Agentic coding is now a team sport

Slack Code brings software development out of private tabs and into dedicated code channels, where the conversation, the code diffs, and the live previews are visible to everyone. Tag in a coding agent — like Claude, Devin, Github Copilot, ChatGPT, or Vercel — from any conversation, and it spins up a channel to tackle the task. Everyone can watch the work happen, give feedback, and sign off before anything ships. No ticket, no meeting, no waiting.

That’s not the only thing new for agent work:

• Agents tab — every agent and agent session gets a home base, so you can browse conversations, check statuses, and pick up where you left off.
• Agent DMs with real titles — threads with agents now work like conversations with any teammate.
• Renamed Tools → Agents & tools — one less thing to hunt for.

Slackbot gets more room to think

Sometimes a question needs more than a quick reply. This month, Slackbot’s capabilities grew to match.

• Big Mode — a full-screen Slackbot workspace built for research, writing, analysis, and file generation, without leaving the conversation.
• Deep Research — ask an open-ended question, and Slackbot digs across your org’s data and the web, reasoning through what it finds instead of just handing you a quick answer.
• Skill Sets — bundle related skills together and deploy them as one coordinated package, instead of piecing them together one at a time.

Your business systems are working even closer together

More of your tools are meeting you where you already work.

• Salesforce approvals in Slack — approvals now route straight into Slack, powered by Salesforce flow orchestration, so approvers see the full picture and decide faster.
• Sales tracker for Slack CRM — teams on the older Contacts and Accounts views are getting a visual pipeline, no setup required.
• Seismic enterprise search — connect Seismic as a search source, and its content surfaces right alongside the context.

Upgrades that make work simpler

We know that little things add up to big improvements, so we’re always focused on the smaller changes that have an outsized impact.

• Comment-only canvas role — canvas owners and editors can share with comment-only access, so reviewers can weigh in without editing. We know you’ve been waiting for this.
• MCP tools for Slack Lists — agents connected through Slack’s MCP Server can now search, create, and update Lists directly.
• Workflow Builder quality improvements — build object collections, filter Lists, and run any step across a group with the new Repeat step, no code required.

Ready to put it all to work?

Check out our Innovations webpage for previous releases.

Please note that some features may not be immediately available in your workspace depending on the rollout timeline, your Slack licensing plan, or additional license requirements. To learn more about feature availability, check with your admin on the licensing plans for your workspace.

Ready to learn more? Contact us.

Slackbot now scales your best workflows, develops cited research, bundles skills as one unit, within a full-screen AI workspace in Slack.

Key takeaways

The hard stuff doesn't need a separate tool anymore:
Big Mode gives Slackbot room to do deep research, multi-step analysis, and real drafting, all without leaving Slack.

One person's setup becomes everyone's default:
Skill Sets let your team bundle related skills and roll them out together in one click, instead of sharing them one at a time.

The answer that used to take hours of digging now comes with reasoned answers:
Deep Research digs across your org's data and the web, then synthesizes what it finds into a report with citations.

Every few months Slack ships something that changes what’s actually possible in the app.

In June, Slackbot learned who you are and connected to your systems. In July, it put that knowledge to work, giving everyday users AI superpowers to move work forward, inside the conversation.

In August, Slackbot starts scaling with your team, not just answering one person at a time.

Skills, work together as a team

A single Skill already saves time. The catch is sharing it: today, that’s a one-at-a-time drag, sending each skill and asking each person to accept it. Skill Sets fixes that. Bundle related skills into one collection and share the whole thing in a single click.

An IT admin building a “Helpdesk AI Kit”, a triage skill, a ticket-creation skill, an escalation skill, shares it org-wide once. Every new hire gets the whole kit on day one, and when the admin adds a new skill to the set, it shows up for everyone automatically. A sales leader can do the same with an “Account Research Kit”: brief, competitive intel, and call prep skills, so the whole team runs the same playbook instead of everyone reinventing their own. And once a team builds something that works, they can share it with a team in another Slack org, without that team having to rebuild it.

It reasons, instead of just retrieving

Deep Research dives into insights across your org’s data and the public web, following the thread wherever it leads, then reasons through what it finds to provide a synthesized report. It’s built for the questions that don’t have a clean answer sitting in one document and that require actually thinking through, not just retrieval.

An account exec prepping for a competitive deal can ask Slackbot for a brief on a prospect’s tech stack. Deep Research pulls from Slack account channels, Salesforce opportunity history, and web research on the prospect’s vendors, and returns a cited report in minutes instead of the hours it usually takes to piece that together by hand.

A workspace built for the deep work

Big Mode turns Slackbot into a full-screen, agentic workspace. Built for long-form research, multi-step analysis, file generation, and drafts that need iteration. You get the space of a dedicated tool with none of the tab-switching, because it’s still the same conversation, just with more room to work.

An account executive planning a trip to meet clients can open Big Mode, run the Account Roadshow skill, and see a Geographical Pipeline surface ranking open deals by state, stage, and account, so they know exactly which meetings to prioritize before they ever leave for the airport.

Ready to see it in action?

Start with Skill Sets, bundle up the skills your team uses most, and deploy them together instead of piecing them together one at a time.

What shipped on August 26

Big Mode.

A full-screen Slackbot workspace built for deeper, longer-form work like research, writing, and analysis, without leaving the conversation. Draft a multi-page strategy brief or post-mortem, then push the finished doc straight into a Surface and pin to a channel for the team to react to in real time.

Skill Sets.

Bundle a related set of skills the whole team can run on, so teams get a coordinated set of capabilities from day one. Assign a “Sales Enablement” or “Incident Response” Skill Set straight to a user group or channel, and everyone gets the same workflows from day one.

Deep Research.

Ask a complex or open-ended question and Slackbot digs across your org’s internal data and the web, reasoning through what it finds instead of just returning an answer. Before a client pitch or exec review, turn years of history from Slack Search, files, and live web data into an executive summary attached right to the thread.

Tasks Tab Enhancements

A dedicated tab to view, run, and manage all your scheduled Slackbot tasks in one place. Project managers can track, edit, and trigger recurring routines, like daily async standups or weekly status pulls, without cluttering team threads.

Skills Admin Dashboard

An admin view showing which Slackbot skills are being used across your org, with usage stats and management controls. Enterprise admins can audit Slackbot Skills usage and permissions, then roll out top performers org-wide to boost adoption and stay compliant.

Slackbot External Skills Share

Skill authors can generate a link so someone in a different Slack org can copy the skill into their own library. For example, agency partners working with vendors in a Slack Connect channel can share a proprietary skill by link, letting the other side clone and adapt it locally without live access to internal systems.

Availability of Advanced Features

Slackbot continues to deliver what we’re calling Advanced Features: capabilities that do more and need more compute to run. Advanced Features will keep showing up wherever Slackbot does.

Quick note on availability: there are no limits to Advanced Features at launch for Enterprise+ customers. Down the road, we may introduce limits given their higher compute costs, but we’ll give customers 30 days notice before that happens. Standard Slackbot messaging use outside of these new Advanced Features remains unlimited for Enterprise+ customers.

For more on Slackbot availability, check out our help center.

We're rolling out a wave of updates to agents on Slack, along with an important heads-up about the road ahead for the Assistant messaging experience.

Agent updates and the deprecation of assistant_view​

The Assistant messaging experience (assistant_view) will be deprecated in February 2027. The assistant-related methods that power it are being replaced by a new set of API methods: the Agent Sessions API. Agent sessions let users manage their conversations with your agent in one place, with a status, a title, and a stop button. The agents.sessions.setStatus and agents.sessions.rename API methods replace the assistant.threads.setStatus and assistant.threads.setTitle API methods, respectively. For now, your existing calls keep working through a compatibility bridge. See the Agent sessions guide to get started.

The Agent messaging experience (agent_view) is already available, and we recommend moving to it now. If your app still uses assistant_view, follow the migration guide to move to the new experience. Read more in the guide to developing agents.

What's new for agents​

Alongside this announcement, we're releasing a set of improvements to the agent experience:

• Upgraded Tools tab: a newly renamed and revamped Agents & Tools tab helps users find popular agents with a new landing page.
• Improved thread experience: including multi-agent stacking, thread naming, a revamped thinking UI, and a native stop button for agent streams. When a user stops your agent, your app receives an agent_session_stopped event.
• Plan and Task card block updates: UI refinements to the plan and task card blocks agents use to show their work.
• Marketplace data in the About tab: an agent's Marketplace data now appears in its About tab.
• Agent sessions in the sidebar: A new sidebar with pin / archive / unread and live updates.

An agent can answer a question in a thread, but some tasks become bigger than a single question. A coding agent may need to implement a change, share a prototype, and iterate on it. A research agent may need to gather information and work through decisions with the people involved. This kind of work needs more than a single response.

Slack Code is how agents and people work together to build with AI in Slack, creating dedicated channels for more involved work.

A channel is created from an existing conversation, context is carried forward, and agents and collaborators have a shared, visible place to work.

Within a session, agents can publish artifacts for people to oversee and iterate on:

• Code diffs, so people can review what changed
• Block Kit views, for structured interactive experiences
• HTML previews, for prototypes and generated interfaces
• Canvases, for longer-form content and collaboration

The work becomes multiplayer: teammates can join channels, review work, add context, and help make decisions, all while collaborating in the same session and with the same artifacts. Your agent manages the lifecycle, creating a channel when a task warrants one, sharing updates as the work progresses, and closing the session once the work is complete.

Read on to learn more:

• Slack Code: where your team and agents build together
• Build with AI as a team using Slack Code

How code channels turn AI from a solo tool into a teammate

Author: Katie Steigman, Claire Bowman

The real work has never been the document, the pull request, or the final build. It’s the conversation that gets you there, the back-and-forth, the pushback, the half-formed idea a teammate catches and turns into something brilliant. A brief, a deck, a block of code: they’re just proof that the real thinking already happened, together.

We’ve believed this since Slack started, that work is better done in the open, in channels, not alone behind a closed door and revealed only once it’s “done.”

AI agents haven’t always been part of that. Usually, someone opens a private tab, prompts an agent, and disappears for a while, building in isolation, one person and their agent. The rest of the team finds out after the fact, with no chance to weigh in, catch a wrong assumption, or add context only they had. It’s a single-player game, and single-player games don’t scale: no pushback, no collision of ideas, just a fast way to get more of the same thinking. It’s the exact thing we’ve spent years building Slack to avoid, now happening again with agents instead of people.

Welcome to Slack Code

Slack Code brings software development out of private tabs and into the open, so your team and its agents can plan, write, and review together. It provides a dedicated space, called a code channel, that shapes itself around the work. Mention a coding agent on a complex project, and one spins up, pulls in anyone on your team, and the agent builds the space around what you need, like a code diff, a planning doc, or a live HTML preview. When the work’s done, the channel archives itself but remains like an audit log in Slack for future reference. That’s the shift Slack Code is built for: connecting what you’re building to the conversation your team is already having, in one place.

We’re starting with engineering because it’s the work closest to our hearts, the thing we live and breathe every single day. Today, we’re launching alongside Anthropic, Cognition, GitHub, OpenAI, and Vercel, with more partners joining us on the way.

"Slack is a strategic part of a broader GitHub promise: humans set direction, agents close the loop. That has to be true wherever developers are working, which is why we're bringing Copilot into the flow of conversation in code channels to help teams turn shared intent into shippable software." — Mario Rodriguez, Chief Product Officer, GitHub

Why we built Slack Code

We didn’t design code channels in a vacuum, we built them out of sheer necessity for our own team. As our engineers leaned harder on coding agents internally, we hit a wall: squeezing complex, multi-turn agent execution into a standard channel thread felt cramped and noisy, but letting developers retreat into private browser tabs meant losing the thing that makes Slack work: everyone working in the open.

We realized agents didn’t need another isolated browser tab or terminal. They needed a seat at the table: a dedicated space that could expand when the project got complex, pull in the right humans at the right moment, and fade away when the job was done. Bringing teammates and agents into that same space has made building feel human, transparent, and genuinely fun, and the impact shows up in the numbers, too. One of our engineers put it simply: bug reports and small UI tweaks used to get filed away and forgotten. Now he spins up a side conversation right where the work’s happening, deals with it on the spot, and gets back to building.

Over 70% of code channels spin up and close within a single day, from idea to merged PR. Engineers get unblocked sooner, reviews need less context, and work just ships faster. That’s because the real value of AI doesn’t come from someone working faster in isolation. It comes from agents and teams building together at the speed of conversation. For the first time, an agent can build alongside your team in real time, with your teammates’ eyes on the work as it happens. It’s not solitary, it’s multiplayer, the exact same way your best work already is.

Built for Code, Not Just Conversation

Threads work great for quick questions and updates. But working with agents to test a fix or catch a bug before it ships, needs more room to work. Slack Code gives agents and engineers that space, right where the coding happens:

• A dedicated space per project.
  Mention an agent on real work, and a code channel spins up around that specific task, visible to the whole team, not just the person who kicked it off. It archives automatically when the work’s done, so nothing sprawls, but everything stays searchable.
• See the code as it’s built.
  Code diffs, and live previews show up right in the code channel, not a wall of text. Catch a bug before it ships, flag a constraint before the agent runs with it, or pause and redirect mid-task.
• Enterprise trust, no extra setup.
  Slack Code inherits Slack’s existing permissions and admin controls, so IT doesn’t need to configure or audit anything new. High-stakes changes can route to a person for a fast approval, right in the channel — the speed of automation without giving up the confidence of a review.

Partnering with our ecosystem to bring the best of coding to Slack

Slack Code is launching with AI partners who’ve built their agents for this new and better way of working — one that fits into the natural ways teams already collaborate.

• Claude (Anthropic):
  When a user asks, Claude Tag can invoke a Code Channel, a home for the work so it doesn’t get lost in a thread and separates it from other chats. Everyone can follow progress at a glance through the summary feature in the original thread, jump in to help steer when they have something to add, and ship changes together with Claude Tag, which later archives the room once the work lands.
• Cognition:
  Cognition built Devin to act like a coworker: it responds when needed and stays quiet otherwise, fitting into how a team already talks instead of adding another channel to check. Someone flags a bug right in the conversation, and Devin fixes it and reports back with a working preview. No ticket, no queue, no context lost in translation. For Cognition’s users, that turns ‘we’ll get to it next sprint’ into same-day fixes.
• GitHub:
  GitHub built its integration to open the door to more contributors. Inside a code channel, a non-technical teammate can describe a problem in plain language, have an agent draft a fix, and tag in an engineer to review it — all without leaving the conversation. That turns code review from something isolated and delayed into something the whole team can follow and weigh in on as it happens.
• Vercel:
  Inside a code channel, Vercel’s agent kicks off a live preview the moment a change ships and posts a shareable link right in the thread — so the team sees the real result before it reaches users. Bugs and backlog items that used to sit unprioritized because they ‘weren’t worth’ a developer’s time now just get fixed.

These examples are just the start, and you can build this way today. Simply mention Claude, Devin, GitHub Copilot, ChatGPT by OpenAI, or Vercel in Slack to watch a code channel spring to life in real time. Software engineering is where we started, but together with our ecosystem, this same collaborative model will expand beyond code, stretching into marketing campaigns, legal contract reviews, IT onboarding, and every corner of your business.

Beyond Slack Code

Slack Code is one piece of a much bigger picture. As agents become part of your daily work across Slack, not just in code channels, but also in DMs, threads, and everyday conversations, we’re building the surfaces to help you manage that work, see it happen, and keep it moving. That starts with the new Agents & Tools tab, your home base for managing agent conversations across threads, DMs, and code channels. It means making a DM with an agent feel as natural as messaging the person sitting next to you. And it means sweating the details, like streamlined agent reasoning, and a clearer way to see when an agent’s still working.

Work Is Multiplayer in Slack

We said it earlier: the real work is the conversation that got you there. For years, we’ve built Slack on one belief: work is better, faster, and more creative when it’s done together, in the open.

As agents take on deeper work, we’re rolling up our sleeves to make sure people stay at the center of it. Our goal isn’t just to make agents run smoothly inside Slack. It’s to shape the future of how teams work in an AI-powered world. We want that future to be multiplayer: one where work is more collaborative, agents are not just tools but teammates, and work is, frankly, more fun.

Agents are going to push work forward. We’re here to make sure they do it the way we always have: together, in the open, side by side.

Availability & next steps

Slack Code is live today for teams using Claude (Anthropic), Devin (Cognition), GitHub (Copilot), and Vercel integrations, with OpenAI (ChatGPT) available soon. Mention the agent in a project channel to see it spin one up.

Explore the announcement.

Build and deploy AI agents in the flow of work, no code required

Slack enables teams to deploy agents in the flow of work, backed by enterprise-grade security and IT control.

What is Add to Slack?

Build with your favorite coding or AI creation platform

Ready to bring your agents into the flow of work?

Key takeaways

• Build with your favorite tools:
  Create an agent on any partner platform, click "Add to Slack," and presto—you're done in a few clicks.
• Work together in channel:
  Deployed agents live directly in your channels, collaborating alongside your team just like another teammate.
• Secure by default:
  Every agent automatically inherits Slack's enterprise permissions and data boundaries, keeping IT in full control.

Build and deploy AI agents in the flow of work, no code required

We’re excited to make it possible for anyone to be a builder on Slack with the launch of Add to Slack, an enterprise-grade deployment path built alongside 10 leading AI creation platforms, including Hyperagent, LangChain, Lovable, n8n, NanoClaw, OpenAI, Runlayer, Skydive, Superhuman, Vercel.

Add to Slack is a secure, one-click bridge embedded inside your favorite agent builder tools. It lets anyone deploy custom agents straight into Slack channels in seconds, with no custom integration code, complex webhooks, or technical setup required.

Every day, teams across Marketing, HR, Sales, and Engineering work in Slack to solve problems and move the business forward. As AI reshapes how that work gets done, non-technical teams are increasingly turning to no-code and low-code platforms to spin up their own agents that can tackle tasks and take on parts of their processes. For IT leaders, that raises a familiar challenge: how do you empower teams to innovate quickly while keeping corporate data and AI tools fully governed?

Add to Slack solves this by establishing a standardized, secure pathway into the workspace. Whether you’re an HR lead automating onboarding, a marketer pulling campaign metrics, or an engineer setting up production infrastructure, you can now bring a specialized AI agent into your daily workflow with a few clicks. Business teams get instant, in-the-flow automation, while IT maintains full visibility, permission controls, and data sovereignty.

What is Add to Slack?

Add to Slack is a secure, one-click bridge that connects your favorite AI creation platforms directly to Slack. Anyone, whether developer or not, building with a coding agent or a point-and-click platform, can deploy a working agent into their team’s workspace in seconds. No custom integration code, complex authentication flows, or platform migration required.

• One-click authorization:
  Build an agent and add it to Slack right from your agent builder tool. The platform handles multi-tenant permission scoping and authentication automatically. Closing the door on shadow AI.
• Built-in governance and context:
  Every agent you deploy automatically inherits your workspace’s permissions and data boundaries, giving IT centralized visibility through the Agent Browser.
• In-the-flow collaboration:
  Teammates can immediately mention, task, and interact with the agent in open channels, turning solo automation tools into multi-user team assets.

Imagine building a customer FAQ agent in Lovable using plain text, no coding required. Click “Add to Slack,” and the agent installs straight into your workspace, ready for your team to ask questions in the channels where they already work. You build where you’re comfortable, and your agent appears exactly where your team works.

Lovable already lets anyone build an agent just by describing it. Now you can deploy that agent straight into Slack, ready to use, without your team having to set anything up or learn a new habit. Because building something great only matters if people actually use it.

Build with your favorite coding or AI creation platform

We’re excited to be launching Add to Slack with 10 creation platforms: Hyperagent, LangChain, Lovable, n8n, NanoClaw, OpenAI, Runlayer, Skydive, Superhuman, Vercel. Connect these through the Slack Marketplace and start building agents right in your flow of work.

Hyperagent

Hyperagent is a platform for building AI agents that run in the background finding leads, building campaigns, and managing operations for you. With “add to Slack,” those agents deploy in a few clicks and no platform migration, so the value they create shows up immediately where you and your team already work, and where they can collaborate with you and each other.

On Hyperagent, autonomous agents run 24/7 in the background. When they need to provide an update or ask for an approval, Slack is the most natural place to bring humans back into the loop. With 'Add to Slack,' that connection is a few clicks away, so customers get from a working agent to a collaborative one in seconds.

Skydive

Skydive lets you hire AI agents that take on real work across your existing tools, picking up engineering tickets and PRs, clearing support queues by digging into logs and billing, and triaging inboxes and calendars. With Add to Slack, install a Skydive agent in a few clicks and start working with it right inside your Slack channel.

Slack is where work actually happens, so it's where agents should live. Add to Slack removes the last hurdle between building an agent and putting it to work, and that's exactly the future we're building Skydive for.

LangChain

LangChain lets teams build agents with natural language and deploy them to Slack in one click. From executive assistants managing calendars and briefs, to GTM agents researching accounts and drafting outbound, to engineering agents triaging issues and summarizing PRs.

Fleet makes it easy for teams to build agents around the work they already do and bring them into Slack. Since Slack is where teams already coordinate, make decisions, and move work forward, employees can collaborate with agents right in the flow of work, with the security and control companies need to manage agents at scale.

Lovable

Lovable lets you turn ideas into software using plain text. Now anyone can create an agent just by describing what they want, and deploy it straight in Slack with the right configuration and permissions, so you can use it where you already work.

n8n

n8n is where teams build custom AI agents and workflows to automate routine work like answering support and ops questions from knowledge bases, triggering lead enrichment or onboarding on request, and posting scheduled digests from Linear or CRMs. With Add to Slack, connecting one to your team takes just a couple of clicks.

Building and publishing an AI agent in n8n is easy, though getting the whole team to actually use it can be challenging. The Add to Slack is what makes it easy with just a couple of clicks. You build the agent in n8n, give it the workflows and integrations you already run your business on, and your team starts talking to it in Slack.

NanoClaw

NanoClaw is an open-source framework for agents you can trust, each running isolated in its own sandbox with its own memory, skills, and persona. With Add to Slack, one message can stand up a whole team of agents — a writer, a dev, a designer — that join your channels, take on distinct roles, and collaborate in a shared Canvas.

We want getting one agent live in Slack to take a few clicks, and getting a whole team of them just as easy. With Add to Slack, a single message stands up a crew of NanoClaw agents, each isolated in its own sandbox, each collaborating right where your team already works. That is how agents go from a demo to daily use.

Runlayer

Runlayer is where enterprises build and run AI agents that work like teammates, each with its own identity, full context, and governed, audited access to company tools — created for whole projects like support triage or on-call investigation. Add to Slack provisions them into a workspace in one click, ready to @mention, DM, and work like anyone else on the team.

At the most AI-forward companies, people and agents are working side by side, and Slack is where that starts. 'Add to Slack' lets our customers put a whole roster of Runlayer agents in front of their company, each with its own identity, permissions, and audit trail. With Runlayer agents, even more work gets done in Slack.

Superhuman (Superhuman Go for Slack)

Superhuman Go lets teams build custom AI agents connected to their knowledge, tools, and workflows, from a prioritization agent surfacing top accounts, to a product agent answering roadmap questions, to a support agent instantly answering from company docs. With Add to Slack, teams interact with these agents alongside teammates without switching apps.

Slack shares our commitment to helping customers use AI to work more effectively. That's why we're excited to bring custom Superhuman Go agents into Slack, making it easier for organizations to collaborate with AI alongside their teams in the conversations where work gets done.

Vercel

Following our Dev Day preview, Vercel’s agent framework lets developers and technical leads spin up production infrastructure, automated monitoring, and deployment bots in Slack in minutes, enabling them to build and bring specialized agents directly into their daily workflows.

Ready to bring your agents into the flow of work?

No custom code. No shadow AI. Just rapid, secure innovation right where your team collaborates. Whether you’re an engineer orchestrating production workflows or an HR lead automating employee onboarding, add to Slack makes it seamless to bring specialized agents straight into your daily conversations.

Ready to transform how your team works? Head to the Slack Marketplace today to explore our launch partners, build your first agent, and deploy it to your workspace in just a few clicks.