10 Years in Hospitality Operations Taught Me What Most AI Companies Get Wrong

Before I started On The Hill AI, I spent over a decade in operations management. I ran transport logistics, managed multi-site retail, coordinated delivery schedules, built staff rotas, and dealt with the daily chaos that anyone in hospitality knows intimately.

I didn't learn about hospitality from a case study. I learned it from standing in a walk-in fridge at 5am checking a delivery that was missing half the order, then spending the next hour on the phone to a supplier while also trying to figure out who was covering a no-show on the morning shift.

That experience is the reason I now build AI automation for hospitality businesses. And it's also the reason I think most AI companies get it fundamentally wrong.

Most AI Companies Automate the Wrong Hospitality Processes

The typical AI consultancy walks into a hospitality business and immediately goes for the shiny stuff. Chatbots on the website. Social media content generators. Customer sentiment dashboards with fancy graphics.

None of that solves the actual problem.

The actual problem is that the manager spent their morning chasing a supplier order — or losing bookings because enquiries sat unanswered overnight that should have been confirmed automatically, then rebuilding the rota because someone called in sick, then manually responding to 15 booking enquiries that came in overnight, then pulling together a weekly report from three different spreadsheets. By 2pm they haven't done any of the work that actually moves the business forward.

When I look at a hospitality operation, I start with the processes that eat the most management hours per week. Not the processes that make the best demo. There's a difference, and most AI companies don't know it because they've never lived through a double shift on a bank holiday.

Why AI Systems Fail in Real Restaurant Environments

Here's something I learned from years of building systems inside operational businesses: the gap between how a system is designed to be used and how it actually gets used at 7am on a Monday by someone who's already stressed is enormous.

A technically brilliant automation that requires three clicks and a confirmation step will get ignored. A slightly less elegant system that fires automatically and only surfaces when something needs human attention will get used every single day.

I've seen expensive technology projects fail not because the tech was bad, but because the people who designed it had never worked the kind of shift where you're simultaneously taking a call, checking a delivery, and trying to remember if you've confirmed tomorrow's bookings. If your automation adds friction to someone's day, they will work around it, and you've just wasted your money.

The POS and Booking System Integration Problem Nobody Mentions

Hospitality businesses don't run on one system. They run on a booking platform, a POS, a supplier portal (or three), a scheduling tool (or more likely a WhatsApp group and a spreadsheet), email, and whatever accounting software someone set up five years ago.

Most AI consultancies want to sell you a new platform. I think that's backwards. You don't need another system — you need the systems you already have to talk to each other and do the repetitive work automatically.

That integration layer is where the real value sits. It's also where most projects get complicated, which is why a proper Discovery process matters more than a flashy demo. I need to understand what you're actually using, how your team actually uses it, and where the real bottlenecks are — not where you think they are.

Realistic AI Automation Results for UK Hospitality SMEs

I see AI agencies claiming 70% cost reduction, 10x productivity gains. Meanwhile, the UK government's Technology Adoption Review confirms that AI delivers measurable but realistic benefits, complete transformation in 30 days. I've been in this industry long enough to know those numbers need a lot of asterisks.

Here's what I'll tell you instead: for most hospitality SMEs, well-targeted automation will save the management team 10–20 hours a week — and here's exactly what it costs, reduce manual errors in ordering and communication, and pay for itself within six months. That's not as exciting as "10x productivity" but it's real, it's measurable, and it's achievable.

And sometimes the honest answer is: you don't need automation yet. If your processes aren't repeatable, if your data is all over the place, if you haven't standardised how things work across sites — you need to fix that first. I'd rather tell you that upfront than sell you a system that papers over a process problem.

Why Hiring an AI Consultant With Operations Experience Matters

I'm not writing this to bash other AI companies. Some of them are good. But I think there's a meaningful difference between someone who's studied hospitality operations and someone who's lived them. I know what it feels like when a supplier lets you down at the worst possible moment. (I built custom automation systems from inside a multi-site London operation to solve exactly these problems.) I know why your managers don't fill in the tracking spreadsheet. I know why your booking confirmations are inconsistent across sites.

That's the context I bring to every automation I build. Not just "can this be automated?" but "will this actually get used by a team that's already stretched thin, and will it make their day better or just add another system they have to check?"

If that sounds like the kind of approach you want, we should talk.

Frequently Asked Questions