Slack president Denise Dresser told TechCrunch that the company is transitioning its business chat platform into an “operating system for work,” specifically, making Slack a hub for AI applications from Salesforce, Adobe, and Anthropic. The company’s CEO sees Slack as more than a place to chat with colleagues, but do users want that? And if they do, will they pay a premium for it?
Slack on Monday announced several new features for Slack AI, the messaging platform’s higher-priced tier. The updates include AI-generated Huddle summaries, similar to the channel summaries already available to subscribers. Users can also now chat with Salesforce’s AI agents in Slack, along with third-party tools that enable AI web search and AI image generation.
Salesforce acquired Slack in 2021, shortly after the messaging platform became a remote work staple for millions of people. Three years later, Salesforce is pivoting in earnest to AI agents, and apparently the momentum has been so strong that the company’s popular messaging service is also focusing on them. Slack CEO Denise Dresser said the platform will play a key role in the transformation because it’s a place where people can naturally interact with AI agents because they’re already chatting during the workday.
“AI represents a new way to experience technology that’s very natural for Slack: conversationally, presenting information and taking actions in the flow of work,” Dresser, who became Slack’s chief executive officer 10 months ago, said in an interview. “There’s probably no better place or product to do that than Slack.”
But why does Slack need AI? Since ChatGPT was released in 2022, many companies have introduced AI features to appear “cutting edge,” even if the integration doesn’t make much sense for the core product. Slack’s addition of an AI agent to its messaging service seems no glaring exception.
Dresser’s justification for the AI agent is that Slack is not just a work messaging platform, but a digital workplace or work OS that “brings all your people and processes together.”
The head of Slack told TechCrunch that every CEO wants AI capabilities, like ways to quickly capture team discussions and tools to surface information buried in databases. These are some of the small ways Slack is trying to lead the company into the AI era, she explained.
One of Slack’s new agents, Agentforce, will enable Salesforce customers to perform on-demand analysis of their business data directly within Slack, and Cohere and Anthropic’s Slack agents offer similar services, as long as you’re paying for their enterprise AI services.
Perplexity has also released an agent for Slack that enables web searches, and the Adobe Express Slack agent lets you create branded content from text prompts within the messaging service.
Klarna’s CEO made waves last month when he announced plans to drop Salesforce and Workday as software providers and replace them with home-grown AI tools. Andreessen Horowitz Partners published a blog post in July predicting a shift in how companies will move away from expensive CRM services and toward home-grown AI solutions. Salesforce CEO Marc Benioff is skeptical of Klarna’s AI solutions and wants to see evidence that the company is actually following through on them.
Asked about the Klarna CEO’s comments, Dresser said enterprise AI solutions need to be reliable and secure, and that Salesforce is trying to ensure both for its customers.
That trust was tested earlier this year when Slack came under fire for using customer data to train its recommendation system by default, according to a portion of the privacy policy discovered by developers on Hacker News. It was later revealed that Slack did use customer data to recommend emojis, but not to train the large-scale language models that underpin Slack AI. The privacy policy did, however, ask users to send the company an email if they don’t want their messages to be used to train Slack’s training data.
Slack maintained then, and continues to maintain, that it doesn’t use customer data to train Slack AI.
“We don’t have any law graduates trained on Slack data,” Chief Product Officer Rob Seaman told TechCrunch in an interview. “To be honest, we had a bit of a hiccup and our website policies were updated, but we could have handled it better. It’s a situation we didn’t want to be in, especially in the age of AI and increased awareness of how data is being used.”
These questions about privacy are only going to become more prevalent as Slack pushes AI further, transforming the service from just messaging to one where AI tools feed information into and out of the platform. As AI becomes just another tool, it’s natural for users to be skeptical.
