Google Professional Cloud Architect Study Guide

Chapter 12. Machine Learning and AI on GCP: Teaching Computers to “Think” (Well, Sort Of)

Not too many years ago, I was in a boardroom in San Francisco, watching what I can only describe as corporate performance art. The CEO was literally standing at the head of this really big board table, leading his executives in what sounded like a protest chant:

“What do we want?”

“AI!” they shouted back.

“When do we want it?”

“Now!”

Then came the question that brought the whole thing to a screeching halt. The CTO, who’d been quietly sipping his coffee in a chair along the wall near me, cleared his throat and asked, “And this AI, it’s going to do what, exactly?”

Silence. The kind of silence that I’m pretty sure was costing someone $500 per hour in that particular conference room.

The CEO slowly sat back in his chair. The CMO suddenly found her phone fascinating. The CFO started furiously scribbling notes that I’m pretty sure just said “AI?” over and over.

Finally, someone piped up: “Whatever our competitors are doing with it!”

And there it was—the perfect encapsulation of where most companies find themselves with AI. They know they need it—I mean, have you been to any sort of leader-oriented conference in the last few years?—they know it’s powerful (just look at ChatGPT, Gemini, Claude, Grok…), but they have absolutely no idea what problems it should solve for them.

I’ve seen this scene play out dozens of times. Sometimes it’s a panicked email at 2 a.m.: “Our biggest competitor just announced ...