Salesforce Just Gave Slack a Brain — And Microsoft a Problem

TL;DR

• Salesforce announced 30 new AI-powered features for Slack, rolling out over the coming months.
• The headline feature: reusable AI-skills that let users define custom Slackbot tasks once and deploy them across multiple contexts.
• Builds on agentic capabilities Salesforce shipped in January, signaling a strategic pivot toward AI-native workflow automation.
• CEO Marc Benioff showcased the updates at a San Francisco event on March 31, 2026 — escalating the workplace AI arms race with Microsoft Teams.

Salesforce Rewrites Slackbot With Custom AI Automation

Salesforce just dropped 30 AI features into Slack, and the most interesting one doesn’t scream at you from the announcement. It’s what the company calls reusable AI-skills — a way for users to teach Slackbot a task once, then apply it everywhere.

CEO Marc Benioff unveiled the overhaul at a San Francisco gathering on March 31, 2026. The rollout will stretch across the coming months, stacking on top of the agentic capabilities Salesforce pushed into Slack back in January.

The reusable skills concept works like this: you define a specific task for Slackbot — summarize meeting notes, flag budget overruns, route customer complaints — and that skill becomes a template. Once created, you can trigger it in different channels, projects, and workflows without rebuilding the logic each time.

According to Salesforce, “perhaps the most notable feature announced Tuesday is what the company calls reusable AI-skills — which allow users to define specific tasks for Slackbot that, once created, can be applied in a variety of different scenarios and contexts.” That’s the bet: turning Slack from a messaging app into a platform where users build their own micro-automations.

The other 29 features weren’t detailed individually, but the sheer number signals Salesforce is flooding the zone. Thirty features in one announcement isn’t a product update — it’s a repositioning.

Why Reusable Skills Matter More Than the Feature Count

Here’s what Salesforce is really doing: it’s handing users the tools to program their own AI assistants without writing code. And that’s a fundamentally different approach than most enterprise AI plays, which still rely on vendors to define what the assistant can do.

The reusable skills model flips that. Instead of waiting for Salesforce to ship a feature you need, you teach Slackbot yourself. Once. Then you deploy it everywhere.

Think of it like creating a keyboard shortcut that works across every app on your machine — except the shortcut is an AI agent that understands context, reads messages, and takes action. You’re not just automating a click. You’re automating judgment.

I’ve watched enterprise AI tools struggle with this exact problem for years. Companies ship assistants that do five things really well and nothing else. Users hit the edge of those five capabilities within a week, then abandon the tool. Salesforce is betting that user-defined skills solve that adoption cliff.

But there’s a catch. Reusable skills only work if users actually create them — and that requires a learning curve most enterprise workers won’t climb unless the payoff is immediate. Salesforce will need to ship templates, examples, and dead-simple UI to make this stick. If defining a skill takes more than two minutes, adoption tanks.

The timing matters too. This isn’t Salesforce experimenting with AI — it’s Salesforce pivoting its entire business model toward AI-native tools. Benioff has spent the last year remaking the company’s product line around agents and automation. These 30 Slack features are another brick in that foundation.

And let’s be clear: this is a direct shot at Microsoft Teams. Microsoft has been layering Copilot into Teams for over a year, and it’s winning enterprise deals on the strength of that integration. Salesforce can’t afford to let Slack fall behind in the AI race. Thirty features in one drop is a signal that the company knows it’s playing catch-up.

Salesforce’s AI-Native Workflow Gamble Intensifies

Zoom out, and this announcement fits into a broader pattern. Salesforce isn’t just adding AI to Slack — it’s rewriting its entire collaboration strategy around the assumption that every workflow will eventually run through an AI agent.

The January rollout of agentic capabilities was the first move. Those features let Slackbot take multi-step actions — booking meetings, updating CRM records, looping in team members — without constant human prompts. The reusable skills update extends that logic: now users can define the actions themselves.

This is Salesforce betting that the future of enterprise software isn’t apps. It’s agents. And if that bet pays off, Slack becomes the command center where workers orchestrate those agents instead of jumping between SaaS tools.

The competitive stakes are brutal. Microsoft Teams reportedly has over 320 million monthly active users, and Microsoft is embedding Copilot across its entire Office suite. Slack’s user base is smaller, and Salesforce doesn’t have the OS-level integration Microsoft enjoys. That means Salesforce has to win on flexibility and customization — exactly what reusable skills are designed to deliver.

Google Workspace is the other threat, though it’s been quieter on the AI assistant front. If Google ships a comparable feature set into Chat, Salesforce’s window to establish Slack as the AI-native collaboration leader narrows fast.

There’s also the question of whether enterprises actually want this much customization. Some IT teams prefer locked-down tools with predictable behavior. Letting every employee define their own AI skills could create a sprawling mess of undocumented automations that break when someone leaves the company. Salesforce will need governance features to prevent that chaos — and those features weren’t mentioned in this announcement.

What to Watch as Salesforce Rolls Out Slack’s AI Overhaul

First, watch adoption metrics. Salesforce will tout the number of skills created and the frequency of reuse. If those numbers stay low six months after launch, it means the feature is too complex or the use cases aren’t compelling enough. High skill creation but low reuse means users are building one-offs instead of reusable templates — which defeats the entire point.

Second, watch for governance and security features. Enterprises won’t adopt user-defined AI skills at scale without controls around what those skills can access and how they’re audited. If Salesforce doesn’t ship admin tools to monitor and manage custom skills, CIOs will block the feature entirely. That’s not speculation — it’s how enterprise IT works.

Third, watch Microsoft’s response. If Teams ships a comparable reusable automation feature within the next quarter, Salesforce’s first-mover advantage evaporates. Microsoft has the distribution and the installed base. Salesforce has the product flexibility. This race will be won on execution speed, and both companies know it.

Source: TechCrunch