Summary: Salesforce made its rebuilt Slackbot generally available on January 13, 2026, phasing access to Business+ and Enterprise+ customers through January and February while admins can configure or restrict access until February 10, 2026. The assistant searches, drafts, schedules, and surfaces Salesforce context within existing permissions, and is positioned further out to orchestrate Agentforce and third-party agents. Its differentiation rests on governance and permissions, while early time savings remain experiential rather than proven business ROI.
The rebuilt Slackbot is now generally available. Salesforce set the date at January 13, 2026, shipping the new version in phases to Business+ and Enterprise+ customers across January and February; Org Owners and Admins keep the option to configure or restrict access up to February 10, 2026. The old bot that sat quietly in the corner of the app is gone. In its place stands “a deeply personal agent for work” — zero setup, zero training, working straight from the conversations, files, and coworker networks you already have, all while honoring your existing permissions.
Why this announcement carries weight: Agentforce 360 has taken its first step off the Dreamforce stage and into an everyday interface. Every vendor is shipping AI agents right now; Salesforce buried the entry point inside the collaboration tool millions of people keep open all day rather than standing up yet another product surface. Salesforce co-founder and Slack CTO Parker Harris put it plainly: “Slackbot isn’t just another copilot or AI assistant. It’s the front door to the Agentic Enterprise, powered by Salesforce.”
What the Rebuilt Slackbot Can Do
The announcement lays out the core capabilities: searching across channels for answers, files, and decisions so you’re caught up on a project; drafting meeting notes, briefs, and canvases that you then shape through conversation; scheduling meetings, surfacing priorities, and setting reminders; drawing customer context from Salesforce data ahead of a meeting; and, further out, orchestrating Agentforce and third-party agents from one conversational interface.
The pilot roster reads like a credible cross-section of the market: Beast Industries, reMarkable, Xero, Mercari, Engine, Slalom — plus Salesforce itself in the role of “customer zero.” Strip the feature set down and one pattern remains: retrieval plus drafting. Slackbot ingests what your teams have already written, then turns out the artifacts — notes, briefs, canvases — that convert scattered conversations into something reusable.
The Real Pitch: Permissions and Governance
Futurum analyst Keith Kirkpatrick nails the framing: what Salesforce sells as the core enterprise AI differentiator is context governance and permissions, not model capability. Slackbot works within the roles and compliance boundaries already in place, interactions remain private, customer data never goes into LLM training, and the models run inside Slack’s virtual private cloud. In regulated industries, that bundle of guarantees clears procurement far more readily than another round of benchmark scores. A deliberate contrast with consumer-grade assistants is built in as well: the agent already knows your teams, your channels, and your working patterns, so there’s no onboarding flow and no cold start.
Early Numbers, and a Skeptic’s Note
Every figure so far is experiential. Sinan, Beast Industries’ head of marketing, claims savings of “at bare minimum, 90 minutes a day”; Engine’s SVP of Ops, Mollie Bodensteiner, puts it at roughly 30 minutes a day, all of it from eliminated context switching; internal Salesforce teams say several hours a week once spent hunting down information have come back. One customer produced a canvas in 17 seconds; another watched a weeks-long analysis collapse into minutes. Slack’s own marketing adds 25% faster task completion and 3x revenue per employee. Then comes Kirkpatrick’s cold water: unless time savings convert into top- or bottom-line impact, AI ROI “will remain muted.” Futurum also notes that Salesforce is leaning on experience-driven early evidence rather than long-term adoption metrics — exactly why the phased rollout deserves scrutiny. Every productivity agent now sits for that same exam.
Head-to-Head with Claude Cowork
The timing was pointed — Anthropic had launched Claude Cowork just one day before. Both products stake everything on one thesis: agents have to live where work already happens. The leverage each side holds is what differs. Anthropic builds out from documents and desktop workflows; Salesforce owns the default screen of enterprise messaging, with CRM data stacked on top. Developers should read Slackbot’s future role as the orchestration layer for Agentforce and third-party agents as a new integration point and distribution channel. The platform angle counts too: should Slack become the control layer where employees invoke agents, the agent directory and the routing rules turn into the new procurement surface. For product teams, the outcome of this fight over the front door of work will decide what enterprise AI contact surfaces look like through 2026.
Signals Worth Tracking
Futurum’s watch list runs four items deep: how consistent the rollout proves, how admin-level access configuration settles, whether Slackbot claims the primary Agentforce interface role, and whether usage matures from summarization and retrieval into coordinated multi-system actions. The final one is where the dividing line sits. Plenty of agents will find things for you and save you time. Only those that reliably trigger scheduled actions across multiple systems have earned the phrase “front door.” Two more signals matter: whether admins leave Slackbot open by default once the February 10 configuration deadline passes, and which plan tiers show sustained usage once the novelty wears off.