By admin 
Slack is poised to fundamentally transform how artificial intelligence assists in software development, aiming to pull AI coding agents out of solitary terminal sessions and into the dynamic, collaborative environment of group chats. The Salesforce-owned messaging behemoth today unveiled Slack Code, a groundbreaking new product designed to seamlessly integrate a suite of prominent AI coding agents—including Anthropic’s Claude Code, Cognition’s Devin, GitHub Copilot, and Vercel’s agent—directly into dedicated Slack channels. This innovative approach empowers entire teams to collectively observe, guide, review, and ultimately ship software, fostering an unprecedented level of transparency and collaboration in the coding process. Slack Code will be accessible across all Slack plans at launch, though users will require their own subscriptions or access to the partner AI agents they wish to utilize.
The core proposition of Slack Code is elegantly simple yet profoundly impactful: traditionally, interactions with AI coding agents have been a solitary affair, confined to individual developers and their chosen AI assistant, with the process largely invisible to the rest of the team. Slack Code redefines this paradigm by making that work "multiplayer." When a user initiates a request by mentioning a coding agent within any conversation, the agent automatically spins up a dedicated, project-specific code channel. Within this channel, all development activities unfold in the open. This includes transparent code diffs, live previews of changes, and a clearly visible, evolving plan laid out in dedicated tabs. Once the task is completed, the channel is archived, leaving behind a comprehensive and searchable audit trail of the entire process.
Rob Seaman, Slack’s interim CEO, articulated this vision during a press briefing ahead of the launch: "One of the things I love about this is that code is no longer the bottleneck," he stated. "Ideas, taste, judgment, craft – those are the things that are the bottleneck, and you’ve effectively extended the population that can contribute ideas, taste, judgment, and craft to anybody that exists in your Slack." This statement underscores Slack’s strategic belief that the true constraints in modern software development lie not in the ability to write code, but in the collective human intelligence and creative input that shapes its direction and quality.
This launch represents a significant moment for Slack, and it offers a revealing glimpse into the evolving landscape of the broader enterprise AI market. The recent surge in AI coding assistance has, to date, primarily been characterized by a focus on individual productivity – developers working in isolation with tools like Claude Code or OpenAI’s Codex within the confines of a terminal window. Slack’s ambitious bet is that the next phase of AI adoption in development will be dominated by whoever can effectively own and optimize the collaborative layer surrounding these powerful agents. This strategic move comes at a critical juncture for its parent company, Salesforce, which is actively seeking to demonstrate the transformative power of AI to reinvigorate its growth narrative.
Slack Code Channels: Orchestrating AI Coding in the Open
Slack’s vision for code channels positions them as dynamic hubs for AI-driven development, drawing parallels to a highly efficient newsroom rather than traditional pair programming. Jeff Wang, president of new enterprise at Cognition, the creators of the Devin coding agent, illustrated this workflow with a live demonstration. He depicted a scenario where a broken feature is reported in an engineering channel. Devin promptly acknowledges the issue with an emoji, responds directly in the thread, initiates an investigation, and subsequently opens a pull request. Crucially, Wang highlighted that Devin intelligently identifies code owners, tagging relevant individuals like Theo into the process. He emphasized that each AI action is performed within its own isolated computing environment, demonstrating Devin using Chrome and its DevTools to rigorously test the functionality of the corrected feature.
From this initial interaction, the workflow seamlessly transitions into a dedicated code channel. Here, the collaborative potential truly shines. Any team member – be it an engineer, a product manager, or a designer – can readily join the conversation, offer input, and contribute to the development process. In Wang’s demonstration, a designer even intervened mid-task by dropping a Figma file into the channel, and the AI agent effortlessly incorporated the design assets without disrupting its workflow. The agent concluded its task by posting the code modifications alongside visual evidence, including screenshots and a recorded demonstration verifying the successful implementation of the feature. This robust verification loop is central to Slack’s value proposition. Unlike agents operating solely on a developer’s local machine, cloud-based agents can generate an irrefutable, auditable record of their work, providing a high degree of confidence in the accuracy and correctness of the code produced. "Scaling things, auditing things, giving it to everybody – that is much easier with these cloud agents form factor than it is with local agents," Wang remarked, emphasizing the advantages of this centralized, transparent approach.
The impact of Slack Code extends beyond just code channels, ushering in a broader reimagining of how AI agents integrate into the Slack ecosystem. The platform is rolling out a suite of new features, including agent direct messages that mimic natural conversations with colleagues, a dedicated "Agents" tab that provides a centralized dashboard for all agent sessions with live status updates and stop controls, and a streamlined "Add to Slack" flow. This flow allows teams to deploy agents from a wide array of platforms, including Lovable, n8n, OpenAI, LangChain, and Airtable, with just a few clicks, automating complex OAuth and configuration processes.
The Shifting Bottleneck: Why Slack Believes Code Writing is No Longer the Primary Constraint
The strategic underpinnings of Slack Code are rooted in the belief that AI has fundamentally inverted the economic model of software development. Historically, the act of writing code was the scarce and most expensive component of the development lifecycle. Slack’s executives contend that AI has now rendered this step significantly cheaper and more accessible. Consequently, the primary constraint has shifted upstream, now residing with human judgment, strategic direction, and creative oversight.
"One of the things I love about this is that code is no longer the bottleneck," Rob Seaman reiterated. "Ideas, taste, judgment, craft – those are the things that are the bottleneck, and you’ve effectively extended the population that can contribute ideas, taste, judgment, and craft to anybody that exists in your Slack." This perspective suggests a future where the speed and quality of software delivery are determined less by coding proficiency and more by the collective intelligence and collaborative capacity of the entire team.
Cognition provided compelling internal data to support these claims of accelerated development velocity. "We’ve seen our internal merged PR count go up 10x in the last few months, versus our headcount has only gone up like 40 percent," Wang reported. He described a workflow where engineers can initiate a Devin task and then immediately pivot to another project, launching multiple agents concurrently. "Soon you have everybody working on like dozens of agents at a time," he added, illustrating the potential for massive productivity gains.
Wang further elaborated that this paradigm shift extends well beyond traditional engineering roles. "Believe it or not, a lot of our bugs are reported by our sales team. They report it in Slack, and then someone who’s technical applies them to fix the bug." Seaman seized upon this example as a microcosm of Slack Code’s overarching thesis: "So much of that stuff never even made its way to a product manager into a backlog because the communication vehicles weren’t there, the motivation wasn’t there, the knowledge that it could actually be fixed so quick wasn’t there – and we’ve effectively knocked all of that down." This highlights Slack Code’s potential to democratize bug reporting and resolution, ensuring that valuable feedback from all departments is efficiently channeled and acted upon.
Wang ventured further into his vision for the future, predicting that repetitive and less engaging tasks, often referred to as "toil work" – such as fixing bugs, optimizing CI/CD pipelines, or addressing security vulnerabilities – will be progressively automated away. What will remain is work that "requires creativity, planning, business logic." He offered a prediction that may initially cause unease among some engineering leaders: while human oversight remains critical for every code merge today, "I suspect maybe in the next year it’s just going to go through automatically." This bold forecast points to a future where AI agents, with human-guided parameters, could autonomously manage significant portions of the software development lifecycle.
Addressing the "AI Slop" Problem Through Transparency
A primary concern regarding the democratization of software creation is the potential for a deluge of what the industry has termed "AI slop" – output that appears plausible but is poorly conceived or executed, generated at scale by inexperienced users. Slack’s executives, however, argue that visibility, rather than contributing to this problem, actually serves as its antidote.
"The multiplayer part is a guard against that, actually, because people can see your work, people can comment on your work," stated Katie Steigman, Slack’s VP of product. She contrasted this with the current opaque model: "If I’m doing God knows what in terminal with an agent, versus being able to do it in a place where people can see my intent and actually change and shape my work – or slap my hand and tell me that’s slop, because that’s real." This emphasizes the power of collective review and immediate feedback in maintaining quality standards.
Steigman, who identifies as a product manager rather than an engineer, shared her personal workflow as a template for this collaborative approach. "When I put PRs up as a product person, I almost always tag in an engineer from my team. I don’t just send a PR and ask for an approval," she explained. "Almost every time, an engineer will say something like, ‘Come on, you can make that a little bit tighter,’ or they’ll actually give it some specific technical guidance, and the agent will take one more rev and produce code that has been touched by an engineer to a certain extent." This highlights how even non-technical users can leverage AI agents, with the assurance that expert human oversight will refine the output.
Seaman framed this argument in broader, more philosophical terms: "I think the moral arc of multiplayer AI bends towards higher quality and less duplication." He cited Shopify as an example, where CEO Tobi Lütke has reportedly advocated for restricting agentic coding to public channels precisely because it "immediately disseminates every single thing that’s happening in the company" and levels the playing field. Nevertheless, skepticism regarding the widespread adoption and tangible benefits of AI agents persists, supported by industry data.
Gartner’s research indicates that over 40 percent of agentic AI projects are projected to be canceled by the end of 2027, citing escalating costs and unclear business value as primary drivers. Similarly, McKinsey’s latest "State of AI" survey revealed that while 62 percent of organizations are experimenting with AI agents, only about a third have begun scaling AI initiatives, and a mere 39 percent report any discernible bottom-line impact. The persistent gap between enthusiasm for AI agents and their demonstrated value remains a defining challenge in the market Slack is actively targeting.
Slack Code’s Security Framework: No Elevated Permissions, No New Identities
For enterprise customers, one of the most critical design decisions within Slack Code revolves around its permissions model. When questioned directly about the potential for agents in code channels to inadvertently leak access across different teams – for instance, making a finance repository visible to the legal department – Seaman was resolute in asserting that agents inherit only the permissions of the human user who initiates their action, and nothing more.
"Everything is done on behalf of the user, using the user’s ACLs, both in Slack and in the systems that they’re connecting to," he emphasized. "There’s no god permissions or bot-level permissions. Within Slack, the agent has access to information that the user has access to, and access to the channels that it’s been added to." This stringent approach ensures that AI agents operate within the existing security boundaries established for individual users, mitigating the risk of unauthorized access or data breaches.
Steigman further clarified that when an agent initiates a code channel, "the only thing that agent gets from the code channel is the context of the conversation" that triggered its involvement. On the execution side, Wang confirmed that Devin operates within isolated sandboxes, employing "minimum viable access" – and notably, includes an optional mode that completely disables internet connectivity. "You’ve heard all these stories about the agent escaping and causing havoc, but we have different security configurations," he acknowledged, addressing common concerns about AI agent security.
This model of "agents as extensions of existing users" presents a distinct advantage over standalone agent platforms, which often require IT departments to provision new service identities and manage a complex web of one-off permissions. It also effectively addresses the "shadow IT" concerns that have plagued agentic tools. Because agents generate standard pull requests into repositories like GitHub, existing release gates and review processes remain fully intact. "It reduces that barrier upfront to get that initial PR up," Steigman noted. "Then the due diligence still happens in GitHub." This ensures that the integration of AI-generated code adheres to established governance and quality control protocols.
Slack Code’s Significance for Salesforce’s High-Stakes AI Turnaround
The introduction of Slack Code arrives at a particularly dynamic and challenging period for its parent company, Salesforce. The company’s stock experienced a notable decline of approximately 18 percent in the year leading up to January, significantly underperforming the Nasdaq index. This performance reflects growing investor skepticism regarding the long-term impact of AI on the demand for traditional enterprise software solutions. Adding to this pressure, in December, OpenAI recruited Slack CEO Denise Dresser to serve as its chief revenue officer, leading to the elevation of Seaman, then Slack’s product chief, to the interim CEO position.
In response to these market pressures, Salesforce has been aggressively positioning Slack as the primary gateway for AI-driven work within the enterprise. In January, the company launched a revamped Slackbot, powered by Anthropic’s Claude. This update was hailed as the fastest-adopted feature in Salesforce’s 27-year history. Slack Code represents an extension of this strategic vision, moving beyond simple question-answering capabilities to actively producing tangible artifacts within Slack itself, including working code, prototypes, and documents.
A significant strategic pivot is also evident in today’s announcement. In mid-2025, Reuters reported that Salesforce had taken steps to "block rival AI firms from accessing Slack data," a move that signaled a defensive posture. In stark contrast, today’s announcement positions Slack as an open platform actively courting these same AI companies as strategic partners. The company intends to open the code channel APIs to all developers, signaling a commitment to fostering an expansive ecosystem. Software engineering is presented as just the initial use case, with plans to extend dedicated agent channels to marketing campaigns and legal document reviews in the near future. The underlying strategic calculus appears to have shifted from safeguarding Slack’s proprietary data to making Slack indispensable as the central venue where AI agents, regardless of their origin, perform their work.
Industry partners have predictably responded with enthusiasm. Malte Ubl, Vercel CTO, commented, "A whole team can gather in one code channel, watch the agent work, steer it together, and ship a preview." Mario Rodriguez, GitHub’s chief product officer, echoed this sentiment, describing Slack as "a strategic part of a broader GitHub promise: humans set direction, agents close the loop." These statements underscore the collaborative potential that Slack Code unlocks.
The Future of AI Coding: Coexistence of Multiplayer Channels and Single-Player Terminals
While Slack is championing the collaborative power of its new code channels, its executives are careful not to declare the demise of the traditional terminal. When asked about the potential obsolescence of tools like Claude Code and Codex, Seaman predicted a nuanced division of labor. "The overwhelming majority of the work is actually going to happen in these multiplayer environments," he stated. "But there’s going to be deep, immersive, intensive, single-player thought work that’s going to happen in terminals." This suggests a future where complex architectural design and deep technical problem-solving will still occur in focused, individual sessions, while routine tasks, bug fixes, and feature development will be increasingly handled in collaborative, open channels.
Seaman further posited that the adoption curve for these new collaborative AI workflows will mirror the trajectory of previous waves of automation. "People are going to open these things at the beginning" – meticulously reviewing every code diff and every step of the AI’s reasoning process – "and then build trust in the system and open it less and less over time." This gradual process of building confidence and familiarity is crucial for widespread adoption.
This strategic wager by Slack extends beyond a mere product launch; it represents a fundamental reimagining of how software development teams operate. McKinsey’s research consistently highlights that organizations achieving substantial value from AI are those that redesign their workflows around AI, rather than attempting to retrofit AI onto existing processes. Slack Code embodies this principle, presenting a workflow redesign packaged as a user-friendly feature, with the explicit aim of making the team, rather than the individual, the primary unit of AI adoption. If successful, Slack, the platform that revolutionized how colleagues communicate, may well redefine where software is created. However, if the initiative falters, the increased transparency could simply mean that everyone is privy to watching the "slop" pile up collectively.
Regardless of the ultimate outcome, one thing is certain: the era of the lone developer whispering to an AI agent in a private terminal tab is drawing to a close. As Wang aptly put it, "The bottlenecks have shifted." The crucial question that Slack Code aims to answer is whether the collective intelligence and collaborative power of the "crowd" will shrink those bottlenecks, or merely amplify the noise.
