May 2026
This week I read an article in The Times featuring Evan Goldberg, founder of NetSuite. It was written during the London launch of NetSuite's AI-powered task automation tools, exactly the kind of moment where you'd expect big claims and confident predictions.
Instead, Goldberg said: "There are a lot of questions where it is just unknown right now how AI is going to play out."
I found that refreshing. You don't hear that kind of honesty from the CEO of a multi-billion dollar software company, especially not in the middle of an AI cycle where most people are speaking in absolutes.
But Goldberg did say one thing with complete confidence.
"The benefit of experience still exists. That accumulation of real-world experience with real businesses, that data is still not out on the internet. The wealth of data that we have about how customers use NetSuite successfully is our own."
That's the part I keep coming back to.
The Interface Is Changing. The System Isn't.
There's no question the shift in how we engage with software is real. You can already see it in tools like Claude, Google Analytics, and Canva. You don't navigate systems anymore. You describe what you want.
"Show me performance across properties." "Post yesterday's revenue." "Explain this variance."
And you get an answer. The interface is starting to disappear.
But the system underneath doesn't disappear. And neither does what's embedded in it.
Why Hospitality Makes This So Visible
In hospitality accounting, this difference is especially clear, because while it looks structured on the surface, the reality is far more complex.
You have a PMS, a POS, an accounting system, a framework like USALI. On paper, it should all fit together.
In practice, it rarely works the same way twice.
Two hotels can run the same systems and still produce completely different financial outputs. Not because the systems are wrong, but because of how they're configured, how they're used, and how decisions get made at the property level. Revenue is categorized differently. Tax logic is applied differently. Timing between systems is handled differently.
None of that is visible in documentation. You only see it once the systems are actually running.
Experience Is Not Documentation
At Omniboost, we've worked on more than ten thousand integrations over the past seven years. Different PMS systems, different POS setups, different accounting platforms, different countries, different tax rules, different operating models.
We didn't come into this with a predefined model of how hospitality accounting should work. Instead we listened, observed, and let every integration teach us something: what correct actually looks like, where things break, what scales and what doesn't.
Over time, those patterns became a standard. Our standard.
That's the distinction I want to be clear about. Documentation gives direction. It tells you how to use the tooling, how to configure the platform, how to get from A to B. That's its purpose, and it serves it well.
But the standard, what "correct" actually means across thousands of real environments, that comes from experience. And at Omniboost, creating that standard is what we do by design.
Why the Stakes Are Higher Now
For years, the gap between documentation and reality was manageable. Finance teams reconciled numbers, adjusted entries, fixed what needed fixing at month-end. It wasn't efficient, but it worked.
Now the stakes are changing.
AI is sitting on top of these data flows. It's answering questions, generating insights, and in some cases starting to take actions. And when that happens, the quality of the foundation becomes critical.
Because AI doesn't know what correct looks like.
It learns from what it sees.
If the data is inconsistent, it doesn't fix that. It learns the inconsistency. And then scales it.
The Part That Doesn't Change
Goldberg is right. AI won't replace systems like ERPs. It will change how we use them. It will remove friction, make systems easier to access, and return answers faster.
But it won't replace the experience embedded in those systems.
And as the interface disappears, you stop seeing what's happening inside the system. You stop catching the small errors. You stop questioning the output the same way.
That's exactly when experience becomes most important. Not least.
Because when you stop interacting with the system, you start relying entirely on what it already knows.
The foundation has to be right before any of that begins.
That's what we've spent seven years building.
