Why the AI wave follows a pattern every experienced tech leader has already lived through

Most organizations are using AI now. That’s not the problem.

The problem is that the decisions they’re making today have a longer shadow than they look like they do in the moment. And most leaders won’t realize that until they’re already three years behind.

I’ve been in technology for over 30 years. I’ve watched this pattern play out before. And right now, with AI, I’m feeling something familiar. Not déjà vu exactly. More like pattern recognition.

Phase 1 always looks like enough

When the Internet arrived for most businesses, Phase 1 was straightforward. Get connected. Use the connection to do things faster. Email instead of fax. Web lookups instead of phone calls. I remember running applications to do a simple package lookup over a dial-up modem like it was a magic trick. Today, you type a tracking number into Google, and the answer is just there. Ubiquitous. Invisible. Expected.

The value was obvious and immediate. That was enough to justify the investment.

But the companies that only optimized for Phase 1 hit a ceiling.

Phase 2 wasn’t about doing the same things faster. It was about a fundamentally different architecture. Suddenly you needed your applications to talk to each other across channels. You needed a promotional engine that didn’t care whether a customer was in a store or on a website. You needed a cart that followed the customer, not the channel.

Companies that hadn’t built toward that reality spent years trying to undo decisions they’d made in Phase 1 when speed felt like the only thing that mattered. I watched it happen. I was in those rooms.

AI is following the same pattern

Right now, most organizations are in Phase 1 of AI. They’re using Copilot, Einstein inside Salesforce, or an AI assistant baked into whatever platform they already run. They’re getting real value. Meetings get summarized. Content gets drafted. Tasks move faster.

That’s not nothing. Phase 1 matters. You have to go through it.

But it’s not where this ends.

Phase 2 is orchestration

As you move into phase 2, you stop thinking about individual agents doing individual tasks and start connecting them. When the AI that helps you create content hands off automatically to the AI that builds your email campaign. When the agent managing your recruiting workflow talks to the agent managing your onboarding workflow. When those systems aren’t just faster versions of what humans used to do, they’re generating something new.

A digital record of how your business thinks and operates.

That’s your AI IP. And it’s going to be more valuable than most leaders realize right now. The organizations building toward it are accumulating something their competitors won’t be able to easily replicate. The ones who aren’t won’t know what they missed until it’s too late to close the gap.

So. Are you ready?

Here’s where the pattern gets uncomfortable.

The companies that came out ahead of the Internet wave weren’t necessarily the ones who invested the most in Phase 1. They were the ones who made Phase 1 decisions with Phase 2 in mind. They asked, “if we need to connect these systems later, does this architecture let us do that? If we need to scale out or pull in external data, have we built a ceiling that will stop us?”

Most companies weren’t asking those questions. They were asking, “what’s the fastest way to get value today?”

All are legitimate questions. But only one of them protects your investment.

You don’t need to have your AI platform strategy figured out today. Most organizations aren’t there yet and that’s fine. What matters is that the infrastructure decisions you’re making right now—the platforms you’re selecting and the architecture you’re building around—have consequences that extend well beyond the problem you’re solving today.

Three years from now, when orchestration is the expectation and your competitors have been building toward it, you’ll either be ready or you’ll be catching up.

The questions that need to be asked

When you evaluate a technology investment right now, are you asking what it enables in 18 months, or just what it solves today? Are the platforms you’re committing to open enough to connect, scale, and flex as your AI use evolves?

If you can’t answer those questions with confidence, that’s worth knowing now. Not three years from now.

Phase 2 is coming. It always does.

---

This is part one of a four-part series on AI, infrastructure (on-premises and cloud), and the decisions that matter most right now. Read Part 2: The Agentic Moment Is Now. Orchestration Is Coming.