Salesforce on Tuesday launched a completely rebuilt version of Slackbot, repositioning the workplace assistant as a fully autonomous AI agent capable of searching enterprise data, drafting documents, and coordinating tasks — placing it in direct competition with Microsoft Copilot and Google Gemini for dominance in workplace AI.
The new Slackbot is now generally available to Business+ and Enterprise+ customers at no additional charge. Built on Anthropic's Claude large language model and a new enterprise search architecture, the product can access Salesforce CRM records, Google Drive files, calendar data, and years of accumulated Slack conversations. The launch represents the most significant product investment Salesforce has made in Slack since acquiring the platform for $27.7 billion in 2021.
From Notification Bot to Enterprise Agent
Salesforce co-founder and Slack CTO Parker Harris did not mince words about how different the new product is from its predecessor. "The old Slackbot was, you know, a little tricycle, and the new Slackbot is like, you know, a Porsche," Harris said in an interview ahead of launch.
The original Slackbot performed basic algorithmic tasks — reminding users to add colleagues to documents, suggesting channel archives, and sending simple notifications. The new version operates on an entirely different technical foundation, combining a large language model with a robust enterprise search engine and connections to third-party data sources.
"If you've ever had that magic experience with AI — Slackbot is really what we're doing in the enterprise, to be this employee super agent that is loved, just like people love using Slack."
Salesforce chose to keep the Slackbot name despite the fundamental overhaul. "People know what Slackbot is, and so we wanted to carry that forward," Harris said.
Why Anthropic's Claude — and What Comes Next
The choice of Claude as the underlying model was driven partly by compliance requirements. Slack operates under FedRAMP Moderate certification for U.S. federal government customers, and Harris said Anthropic was "the only provider that could give us a compliant LLM" when development began.
That exclusivity will not hold. Harris confirmed that Google's Gemini is coming to Slackbot this year, citing strong performance and cost economics, and said OpenAI remains a possibility. His view of the LLM market echoes that of Salesforce CEO Marc Benioff: "You've heard Marc talk about LLMs being commodities, that they're democratized. I call them CPUs."
On the sensitive question of training data, Harris was unequivocal: Salesforce does not train any models on customer data, citing the structural impossibility of enforcing per-user permissions inside a shared model.
80,000 Employees, 96% Satisfaction
Salesforce tested the rebuilt Slackbot across its entire 80,000-person workforce before launch. According to Slack CMO Ryan Gavin, two-thirds of employees tried the product, with 80% of those users continuing to use it regularly. Internal satisfaction reached 96% — the highest recorded for any AI feature Slack has shipped, according to the company.
Adoption spread largely without top-down mandates. Within five days of internal rollout, employees had organically compiled a shared document called "The Most Stealable Slackbot Prompts," which has since grown to more than 250 entries. Slack principal UX researcher Kate Crotty found that 73% of internal adoption was driven by social sharing rather than management directives.
Employees reported time savings ranging from two to 20 hours per week, according to Salesforce's internal data.
How the Product Actually Works
In a product demonstration, Slack experience designer Amy Bauer showed Slackbot synthesizing qualitative customer feedback alongside a quantitative usage dashboard image, querying Salesforce for enterprise accounts with open deals, generating a written plan in Slack's Canvas document format, and identifying calendar availability across stakeholders — all within a single conversation thread.
Slack Chief Product Officer Rob Seaman described the Canvas creation as a signal of the product's direction. "This is making a tool call internally to Slack Canvas to actually write, effectively, a shared document. But it signals where we're going with Slackbot — we're eventually going to be adding in additional third-party tool calls."
Meeting booking is not available at launch but is "coming a few weeks after," according to Seaman. Mobile availability launches by March 3. Image generation is under consideration but not yet on the near-term roadmap.
Early Customers Report Tangible Time Savings
Among the pilot customers is Beast Industries, the parent company of YouTube creator MrBeast. CIO Luis Madrigal said the security review — often a prolonged process for enterprise AI deployments — moved unusually fast because Slackbot respects existing user permissions, accessing only information each individual already has the right to view.
One Beast Industries employee reported saving a minimum of 90 minutes per day. Other pilot customers include Slalom, reMarkable, Xero, Mercari, and Engine. Engine SVP of Operations Mollie Bodensteiner estimated the tool saves her 30 minutes daily by eliminating context switching.
The 'Super Agent' Vision — and Its Limits
Salesforce positions Slackbot as the central hub in what Harris calls an emerging "agentic enterprise" — a workplace where software agents handle complex, multi-step tasks alongside human employees. Harris envisions Slackbot eventually becoming an MCP (Model Context Protocol) client, coordinating tools and third-party agents across an organization's software stack.
Third-party agent development on Slack is already accelerating. Anthropic last month released a preview of Claude Code for Slack; OpenAI, Google, and Vercel have also built agents for the platform. "Most of the net-new apps being deployed to Slack are agents," Seaman noted.
Harris, however, cautioned against over-promising on multi-agent coordination. "I still think we're in the single-agent world," he said. "FY26 is going to be the year where we start to see more coordination. But we're going to do it with customer success in mind, and not demonstrate and talk about, like, 'I've got 1,000 agents working together,' because I think that's unrealistic."
A Pricing Asterisk
While Slackbot itself carries no additional fee, some enterprise customers could face indirect cost pressures. Fivetran CEO George Fraser has publicly warned that Salesforce's shift in API access pricing could force enterprises to route data through Salesforce Data Cloud or Agentforce rather than third-party tools. Salesforce has described the pricing change as standard industry practice.
What This Means
Salesforce is betting that the tens of millions of workers already inside Slack every day represent a structural advantage no rival can easily replicate — and that embedding a capable AI agent directly into that daily workflow is a more defensible position than asking employees to adopt yet another standalone AI tool.