The enterprise AI boom is a house of cards built on an untested foundation, where obsession with capability has overshadowed unproven reliability under strain. Companies are racing to hardwire generative AI into their core, without asking if it can hold up under pressure.

Mitch Lieberman, VP of Product, Fuel CX at TELUS Digital, sees the enterprise world flying blind toward a predictable failure. He warns this is happening even as consumer adoption of AI skyrockets.

Stress test pending: "What hasn't happened yet is a Black Friday that's depending on LLMs," Lieberman says. "When you have a Black Friday event, everybody makes sure their systems can scale. But now organizations are just asking for responses from foundation models, and I don't think anybody has any concept of the load that's going to be put on them. That will be a very interesting test."

What's the game plan?: Hype, more than strategy, is driving this rush toward a breaking point. Lieberman sees a major shift in focus, as the "why" for adopting AI moves from operational efficiency to revenue generation. But he questions the thinking behind it. "I get the sense people want it because Harvard Business Review said so—just to tell their golf buddies, 'I’m using AI for this thing,'" he says. "You have this great, powerful tool, but what are you doing with it? The answer is often, 'Well, I don't know, but we should have it.' It’s like being told you can hire 100 more people, but having no idea what you want them to do."