A few years ago, electric vehicles moved from niche to inevitability.
Capital flowed. New manufacturers appeared almost overnight. Every pitch deck carried the same confidence. The future had arrived, and incumbents were finished.
Some companies built real engineering depth, supply chains, distribution, service infrastructure. Many assembled narrative faster than capacity. The technology itself did not fail. The field narrowed.
We are watching a similar moment with artificial intelligence.
In commercial real estate and business finance, the language has shifted quickly. AI underwriting. AI outreach. AI deal screening. AI positioning. The suggestion is that anything not wrapped in automation is already behind.
I am not skeptical of the technology. I use it. It improves research, compresses drafting time, and helps organize information at scale. In the right hands, it is leverage.
What concerns me is the reflex to embed it everywhere, even when it doesn’t strengthen the underlying work.
In my Trust as Infrastructure piece, I wrote about the gap between what technology promises and what it earns through repetition and accountability over time. Infrastructure is quiet. It does not announce itself. It proves itself through performance across cycles.
In transactions, trust functions the same way. It is built through pattern recognition, which comes only from exposure. Through sitting across from sponsors who are tired, overextended, or overconfident. Through understanding lender temperament when markets tighten. Through knowing when a deal technically works, and still should not be done.
Artificial intelligence can assist with analysis. It does not assume responsibility for outcomes. It does not absorb reputational risk. It does not repair fractured partnerships when capital structures strain relationships.
During the EV surge, the companies that endured were not those that talked most about electrification. They were the ones who built durable systems around it.
The professionals who endure this AI cycle will be similar. They will integrate the tools without surrendering discernment. They will resist the urge to label every workflow “AI-powered” simply because it can be. They will allow technology to support judgment rather than substitute for it.
The technology will remain and the noise will recede.
What persists is trust built slowly, tested under pressure, and reinforced by accountability.
That is infrastructure.
