Dreamforce 2026 Slack Keynote: Make AI Multiplayer
After watching the Slack keynote at Dreamforce, one thing is clear: the vision is ambitious. Slack doesn’t want to be another place where people ask an AI a question in isolation. It wants to be the front door to Salesforce and the place where people, data, apps, and AI agents work together.
Slack framed this as a new operating system for AI work. The underlying idea is simple: most AI work is still “single player.” Someone sits in a chat window or terminal, gets a useful answer, and maybe becomes more productive. But the learning often stays with them.
Slack’s bet is that when AI work happens in channels, the whole team gets smarter. Prompts, decisions, outputs, feedback, and context become shared, and they compound over time. Slack called this the combination of conversational context and a collaborative interface: the two missing bookends of the AI stack.
Growth momentum and adoption at scale
Slack opened with some big growth numbers: more than 1 billion messages per day, 100,000 new customers this year, 1.5 million weekly users of its MCP server, and 1.5 million monthly Slackbot users. It also positioned Slack as the fastest-growing product inside Salesforce.
The keynote leaned heavily on the idea that leading AI companies already run their businesses in Slack, from Anthropic and OpenAI to Vercel, Figma, Linear, Notion, Granola, and others. The strongest proof point came from Anthropic’s Boris Cherny, creator of Claude Code, who said roughly 90% of his coding happens in Slack.
The reason is not just convenience. It is visibility. When work happens in a shared channel, anyone with useful context can contribute, even if they are not the designated engineer, designer, or product manager.
Three product bets: Slackforce, Slackbot, and Slack Code
Slack is investing in three major areas.
1. Slackforce: Slack as the front door to Salesforce
Slackforce is the vision for doing CRM work conversationally, without constantly switching into Salesforce tabs.
The keynote highlighted three pieces:
- Slack CRM, positioned as native, AI-first CRM inside Slack.
- Slackbot as an MCP client, allowing Slackbot to interact with Salesforce systems and workflows.
- Slackforce Surfaces, dynamic and collaborative interfaces created from a plain-language request.
The Surfaces demo stood out. Instead of assembling a status view from docs, dashboards, and account records, a team could ask Slackbot to build a live workspace for account health, launch feedback, deal coordination, or another recurring workflow. These surfaces are connected to enterprise data, collaborative by default, and usable without development work.
2. Slackbot: AI for every employee
Slackbot is evolving from a personal assistant into something more operational: able to synthesize context across Slack and connected systems, take action, generate artifacts, and participate in shared channels.
Slack previewed rich artifact generation, including presentations and video, along with two-way voice and Slackbot in channels.
The most useful customer takeaway is not “ask Slackbot everything.” It is to identify the repeatable work where context is scattered: meeting prep, account briefs, handoffs, launch readiness, customer escalations, and follow-ups. That is where a connected assistant can save the most time.
3. Slack Code: making building collaborative
Slack Code brings coding agents into dedicated code channels, where technical and non-technical teammates can work together. An agent can build, preview, respond to feedback, generate a PR, and document the work, all in a shared space.
The big promise here is broader than engineering productivity. Slack is trying to make building accessible to anyone with an idea, while still keeping engineers in the review loop.
What customers should do next
- Start with shared workflows, not generic AI adoption. Pick one high-frequency process that already happens in Slack.
- Move more work into open channels. Shared context is the fuel for the model Slack is describing.
- Connect the systems your teams rely on. Slackbot becomes more valuable when it can work across CRM, docs, project tools, and internal knowledge. (If you need help connecting Slack and Salesforce, that is where Atrium comes in.)
- Create a small frontier group. Test advanced workflows with a few motivated people, then bring what works back to the broader organization.
- Make learning visible. The real advantage comes when one person’s AI workflow becomes a reusable team pattern.
Slack’s core thesis is that AI should not just make individuals faster. It should make organizations smarter together. That is the opportunity behind the keynote, and the standard Slack is now setting for itself and its customers. For more on where the Salesforce platform is headed, read our take on Salesforce Headless 360 and the agentic era.
