Salesforce Makes Slackbot An Enterprise AI Agent

Slackbot Moves From Reminders To Enterprise AI

Salesforce has rebuilt Slackbot as an AI agent for workplace users, moving the assistant beyond reminders and notifications.

The new Slackbot is generally available to Business+ and Enterprise+ customers.

The product can search enterprise data, draft documents and take action on behalf of employees.

Salesforce co-founder and Slack chief technology officer Parker Harris described it as the front door to the agentic enterprise.

The change turns Slackbot from a familiar workplace utility into a test of whether employees will let an AI agent operate across the collaboration data they already use.

Salesforce named records from its own platform, Drive documents from Google, calendar information and Slack message history as parts of the data layer.

Claude Powers The First Version

The new Slackbot runs on Anthropic’s Claude model.

Slack’s commercial service operates under FedRAMP Moderate certification, and Salesforce said Claude fit the compliance requirements for the launch.

The rebuilt assistant uses a large language model, search capabilities and connections to enterprise data.

Harris said the old Slackbot was algorithmic and simpler, while the new version is based around an LLM and a more robust search engine.

Salesforce kept the Slackbot name because users already know it.

The product shift is larger than the branding suggests: the assistant is now positioned as a way to bring Salesforce’s AI strategy into daily workplace conversations rather than a separate app that employees must open.

The older Slackbot handled basic algorithmic work such as reminders, channel-archive suggestions and simple notifications.

The rebuilt version is meant to search, draft and act across workplace systems, making the product less of a helper bot and more of an access layer for enterprise data.

Admin Controls Decide The Enterprise Test

For enterprise buyers, the product question is less about whether Slackbot can answer a prompt and more about who controls its access.

An assistant that can read records, files, calendars and message history has to fit existing data policies, user permissions and audit expectations.

The source names Business+ and Enterprise+ customers as the starting audience.

It also places Slackbot inside Salesforce’s broader competition with Microsoft and Google in workplace AI.

Microsoft has Copilot across Microsoft 365, while Google has Gemini across Workspace.

Salesforce is betting that Slack’s conversation layer can become the work surface for agents.

That positioning gives IT teams a practical governance question.

If Slack becomes the interface for actions across records, files and calendars, administrators need clear permission boundaries before the assistant becomes part of routine work.

The launch describes the data connections, but it does not provide customer-level audit results.

Slackbot still needs adoption proof inside customer workflows.

Salesforce described the assistant’s capabilities and model choice, but the launch details do not include customer-level productivity results, cost savings or measured support-load reductions.

Salesforce has made Slackbot more than a notification tool.

The unresolved enterprise issue is whether admins can give the agent enough data access to be useful while keeping permission, compliance and audit controls tight enough for daily workplace use.