For many modern companies, Slack became the accidental connective tissue, almost like a skeletal system holding up the business. It's not just for chat. It became this new layer for automation, a central clearinghouse where data and signals are fed into other AI systems. And now, the rug is being pulled out from under them.
Tools like Slack are the lifeblood of most company communications; the data living amongst its channels is a vital organ to how companies understand their operations. When Salesforce pulled back on Slack's API, banning long-term data indexing and its use for AI training, the swift action cut off key infrastructure supporting companies at large. The move signals the end of the AI era's open playground, where access to a company's own operational life force is no longer a guarantee.
We spoke with Scott Jones, VP of Product at pioneering Vision AI company Realeyes. As an AI product leader with over two decades of experience, he says the change was bound to happen as behemoth platforms woke up to their foundational place in customers' lives.
Connective tissue: For many companies, the Slack API change was a wake-up call to the data they used daily and took for granted. “For many modern companies, Slack became the accidental connective tissue, almost like a skeletal system holding up the business,” Jones says. “It's not just for chat. It became this new layer for automation, a central clearinghouse where data and signals are fed into other AI systems. And now, the rug is being pulled out from under them.”
Just like Google: The seemingly innocuous move is a calculated play straight from the classic platform playbook, according to Jones. “This feels very 'Google-y' to me. They make you very comfortable and then they start putting the vice on you,” he explains. “They’ll tell a story about security and compliance, but they're clearly going to monetize it. It’s the classic model: get users hooked on a free, indispensable tool, and once they’re dependent, change the terms. I’m the only dealer in town, so what are you going to do?”
That playbook is only now possible because proprietary data has become the finite fuel that models desperately need. “The big players have already scraped the internet; what more can you train them on?” Jones asks. “We're back to this idea that data is king, and that the 'free' data that got the AI renaissance off the ground—much of which was effectively stolen—is no longer available. That makes the data inside ecosystems like Slack immensely strategic.”
The classic analogy for technology is you're flying the plane while trying to change the parts without crashing it. Salesforce just did something that could crash the plane for customers who assumed that Slack jet engine was always going to be there for them.
First shot fired: Jones sees this as a bellwether event that will have a chilling effect on how companies innovate and choose tools. “This is very much the first shot fired. Now, everyone is looking around and asking, ‘Who else can I trust? Who else made me feel open and safe, but I’m not?’” he says.
Now developers and startups looking to scale are shifting their postures to risk management. “Every product and engineering team is now going to ask, 'That's a cool solution, but what's our plan if the platform takes it away?' It's going to become a mess of redundancies and hedging."
Flying blind: Widespread uncertainty endangers companies that built their operations on the assumption that vital platforms would remain open. Jones compares it to performing maintenance on a plane mid-flight. “The classic analogy for technology is you're flying the plane while trying to change the parts without crashing it,” he says. “Salesforce just did something that could crash the plane for customers who assumed that Slack jet engine was always going to be there for them.”
