Salesforce Just Gave Slackbot a Porsche Engine. Here’s What That Actually Means.

Salesforce finally did what Slack power users have been begging for for years: they rebuilt Slackbot from the ground up. Not a facelift, not a feature update — a full architectural rewrite. The old Slackbot was a glorified reminder bot. The new one is an AI agent that can search your company’s Salesforce records, Google Drive files, calendar data, and a decade of Slack conversations, then draft documents and take actions on your behalf.

Parker Harris, Salesforce co-founder and Slack’s CTO, put it bluntly in an interview: “The old Slackbot was, you know, a little tricycle, and the new Slackbot is like, you know, a Porsche.” I’ve been using Slack since 2014, and honestly, that’s about right. The old bot was fine for nudging you to add a doc link. It was never going to change how you work.

This new version, now generally available to Business+ and Enterprise+ customers, runs on Anthropic’s Claude. That’s an interesting choice. Harris said Anthropic was “the only provider that could give us a compliant LLM” for FedRAMP Moderate certification, which Slack needs for federal government customers. But he made it clear this won’t stay exclusive: “We are, this year, going to support additional providers. We have a great relationship with Google. Gemini is incredible — performance is great, cost is great.” OpenAI is also on the table.

Harris echoed Marc Benioff’s line about LLMs becoming commodities. “I call them CPUs,” he said. That’s a bit of a simplification — model quality still varies significantly — but the direction is clear. Salesforce wants to be model-agnostic, letting customers pick their preferred engine. That’s smart, because the real value isn’t in the model; it’s in the integration layer that connects the model to your actual work data.

On the training data question — which is always the elephant in the room with enterprise AI — Harris was refreshingly direct. “Models don’t have any sort of security,” he said. “If we trained it on some confidential conversation that you and I have, I don’t want Carolyn to know — if I train it into the LLM, there is no way for me to say you get to see the answer, but Carolyn doesn’t.” Salesforce does not train any models on customer data. Period. That’s the right call, and it’s the only defensible position for an enterprise platform.

Salesforce has been testing this internally with all 80,000 employees, and the numbers are impressive. Two-thirds of employees tried it. 80% of those kept using it regularly. 96% satisfaction — the highest for any AI feature Slack has ever shipped. Employees report saving between two and twenty hours per week. That’s a wide range, but even the low end is significant.

What’s more telling is how adoption happened. According to Slack’s CMO Ryan Gavin, within five days employees had created a shared Canvas called “The Most Stealable Slackbot Prompts.” It now has over 250 prompts, all contributed organically. Kate Crotty, a principal UX researcher, found that 73% of internal adoption was driven by social sharing rather than top-down mandates. That’s the kind of organic traction you can’t fake.

This launch is Salesforce’s most aggressive move yet to position Slack at the center of what everyone’s calling “agentic AI” — software agents that work alongside humans rather than just answering questions. Slack has the advantage of being where people already communicate. Microsoft has Teams and Copilot. Google has Workspace and Gemini. The battleground is which platform can make AI feel like a natural extension of your workflow rather than a separate tool you have to remember to use.

Slackbot’s new capabilities are genuinely useful. During a demo, product experience designer Amy Bauer showed how Slackbot can synthesize customer feedback from a pilot program, pull in a usage dashboard image, and correlate data points across sources — turning scattered enterprise data into something an executive could actually act on. That’s the kind of thing that saves time not in minutes, but in meetings you no longer need.

The question is whether this is enough. Microsoft has distribution on its side — Teams comes with Office 365. Google has the same advantage with Workspace. Slack has to win on quality and integration depth. This rebuild is a strong step. The internal numbers suggest real product-market fit. But enterprise adoption is slow, and switching costs are high.

For now, if you’re already on Slack Business+ or Enterprise+, this is worth turning on. The old Slackbot was a tricycle. This one might actually be a Porsche. Just don’t expect it to drive itself — yet.