March 2026
Over the past few months, I’ve been paying close attention to how tools like Claude are changing the way people build software. And I'll be honest — I'm genuinely excited about what's happening. Not in a cautious, wait-and-see way. I believe this is one of the most significant shifts in how humans and technology interact that I've seen in my career.
What stands out isn't just that AI can write code. It's how it changes the way you think about software entirely.
You no longer have to understand the system in detail. You don’t need to navigate menus or learn workflows. You just describe what you want, and something gets built or executed. That shift might seem small, but it fundamentally changes the relationship between people and software.
At some point, you stop “using” software in the traditional sense. You start instructing it.
For years, software has been built around interfaces. You log in, click through, export reports, adjust entries, and reconcile numbers. The interface was the product. It was also where a lot of validation happened. You saw the data, you questioned it, and you corrected it when something didn’t feel right.
As that layer evolves, something else takes its place.
When you no longer interact with the system step by step, you also stop catching mistakes along the way. You don’t see the misclassified revenue, or the tax logic that doesn’t quite line up, or the journal that needs to be adjusted. You ask for an outcome, and you get one.
That’s where things get interesting.
Because if the foundation underneath that outcome isn’t right, the result isn’t just slightly off. It’s wrong. And you’re much less likely to notice.
There’s a lot of excitement right now about AI in hospitality — faster pricing decisions, better forecasting, more automated operations. I share that excitement. All of it has real potential, and I think we're only at the beginning of what's possible. But too little of that conversation touches on the quality of the data those decisions are built on.
In hospitality, that’s not a trivial issue. Revenue doesn’t come from a single source. It flows through rooms, outlets, packages, and services. Tax treatment varies across markets. Even within the same system, different properties can structure things differently. The same transaction can be interpreted in multiple ways depending on how the system is configured.
That complexity doesn’t go away when you introduce AI. It becomes more important.
If anything, automation and AI increase the need for discipline. When fewer people are involved in the process, and fewer manual checks exist, the system has to be correct by design. Not mostly correct. Correct from the start.
That changes what the product really is.
Here's what I don't believe: that software is going away. What I do believe is that the way we work with it is changing completely. The interface is no longer the most important part. What matters is what happens underneath. The structure of the data. The logic applied to it. The consistency across systems.
Software isn't disappearing — it's becoming invisible infrastructure. And invisible infrastructure needs to be bulletproof.
It’s tempting to focus on what’s new. AI capabilities, new interfaces, new ways of interacting with systems. The possibilities genuinely energize me. But the order hasn’t changed. It still starts with structure. Then automation. Then intelligence.
Skip the first step, and the rest becomes fragile.
The companies that will do well in this next phase won’t necessarily be the ones that move the fastest. They’ll be the ones that took the time to get the foundation right. The ones whose systems produce results that can be trusted, even when no one is watching.
Because when the interface fades into the background and AI does the work, trust becomes the only thing that matters.
Trust is the new interface. And it's earned before the AI ever runs its first query.
