I’ve had some version of the same conversation many times with colleagues and clients over the past two years. They’re all asking the same question, in slightly different words: should we be doing something with AI, and if so, what?

These are smart, experienced leaders who’ve been running serious industrial businesses for decades. They’ve survived recessions, supply chain crises, digital transformation initiatives that consumed years and delivered mixed results. They are not naive about hype cycles. And they’re telling me, with a kind of uncomfortable honesty, that this one feels different. They just don’t know why, or what to do about it.

I want to answer that.

If you’ve been in industrial manufacturing long enough, you’ve watched two or three major technology waves move through your industry: ERP systems, CRM platforms, digital transformation. The story arc was always roughly the same. Early adopters spent a lot of money on immature technology, suffered through painful implementations, and got modest results. To this day, I rarely see a CRM or marketing automation platform working at full potential, and let’s be honest, most of the time they’re disconnected cost centers. Historically, the companies that waited five years, bought the proven version, and implemented it on mature infrastructure did just as well, sometimes better, at a fraction of the cost.

That experience taught a rational lesson: when it comes to technology in our sector, patience is often a competitive advantage. I understand why that logic still feels right. But AI is different, and it doesn’t apply this time.

The previous waves were fundamentally about tools. A CRM system is a tool. Its value sits in the software, and software improves and commoditizes over time. The company that implemented Salesforce in 2005 and the company that implemented it in 2015 ended up with roughly equivalent capability, because the tool itself was available to everyone at roughly equivalent cost.

Commercial AI is not a tool story. It’s a data and capability story, and data and capability compound in a way tools don’t. The company that starts analyzing its warranty and quality claims today will have two years of failure-pattern data by the time its competitor starts. The company that trains an AI system on its customer renewal and churn signals today will have a retention engine that reflects two years of its own account history, its own products, its own market. The tools and software you buy will commoditize. The institutional knowledge built on top of them will not. That asymmetry is new, and it’s why the window to build a compounding advantage is open right now in a way it wasn’t during the previous technology waves. It will not stay open indefinitely.

Almost every industrial leader I talk to is carrying scar tissue from a previous technology initiative. The ERP implementation that took three years and cost twice the budget. The CRM that nobody actually uses. I’ve lived through most of these myself. That reaction makes sense. Technology-driven change is painful, expensive, and more often than not, it doesn’t deliver what was promised. That’s an accurate read of what happened. Where it goes wrong is in blaming the technology.

The ERP didn’t fail because ERP is bad. It failed because the process it was built to support was never well defined, and got harder to adjust as the technology moved faster than the business could keep up. The CRM sits unused for much the same reason: it’s a pain to keep updated, hard to put good information into, and gives almost nothing back on a daily basis. The digital transformation programs of the last decade never really had a shot, because they were designed to change the organization from the outside in, and industrial organizations don’t change that way. Their technology stack doesn’t connect that way either.

I’ve spent fifteen years inside organizations like yours watching these patterns repeat, hoping and planning for something different, and it’s the reason I’m in this business now.

The industrial companies that define the competitive landscape in your sector ten years from now are being decided in the next two or three years, and most of the people making that decision don’t know they’re making it yet. The ones who start building commercial intelligence capability now will have a compounding advantage in quote responsiveness, customer insight, pricing intelligence, and sales productivity, built from more data, better-trained systems, and organizational capability that takes years to build and can’t be purchased off a shelf.

The ones who wait will still be viable businesses. But they’ll be playing catch-up in their commercial functions the same way the laggards in lean manufacturing spent the 1990s and 2000s playing catch-up on the plant floor. Some of them caught up. Many of them never fully did.

That’s the piece that’s genuinely new here, and it’s why the patience that’s served this industry for thirty years is exactly the wrong instinct this time.

Ask one person on your commercial team a simple question: what’s the thing you do every single week that feels like busywork, that you’ve just accepted as part of the job because nobody’s ever fixed it? Write down the answer. Don’t try to fix it yet. Just notice that it exists, and notice how normal it feels.

That answer is where the rest of this series is going. Part 2 picks up exactly there: why the instinct to clean up that process before AI touches it is exactly backwards, and where to begin instead.