AI in hospitality: build the foundations before you chase the future

Arik Fletcher, Head of Strategic Consulting, Focus on Hospitality by Focus Group, recalls the discussion around AI at the Next Wave Hospitality event in London earlier this year.

Start with the problem, not the solution

The question I keep coming back to in almost every conversation I have with hotels and operators is: what problem are you actually trying to solve? It sounds almost too simple, and yet it is the question most businesses skip.

AI is just a tool. What problem are you trying to solve?

Every industry is asking this right now, and hospitality is no different. The honest truth is that AI is not going to fix a broken process, a poorly-structured dataset, or a culture that does not yet understand why technology investment matters. AI is a force multiplier. Bad data, bad systems, and bad practices will get worse with AI. Good practices, good processes, and good systems will get meaningfully better.

The reason to move is the same as the reason to move carefully: deliberate action compounds faster than reactive action.

The infrastructure conversation nobody wants to have

We talk a lot about AI capabilities, but we do not talk enough about what needs to be in place before any of it can deliver value. Infrastructure is the unglamorous, non-negotiable foundation. You cannot avoid it.

In a typical hotel environment you might have 20 or more applications, and right now every single one of those vendors is building some version of AI into their product. Within a year or two, we risk having 20 AI agents running in parallel, each operating on its own slice of data, and ending up back where we started: siloed information, conflicting recommendations, and no single source of truth.

We have tried to solve the data problem in hospitality for decades. AI gives us a genuine opportunity to do it properly. But only if we architect it that way from the start.

The organisations I see having the most success are those that were already keeping pace with technology before AI came along. They do not have the legacy burden to carry. If you are starting from scratch, you can build it right. If you have kept your systems current, you are in a strong position. But if your files are still sitting in a back room somewhere and your systems are monolithic and disconnected, AI is simply not going to work for you yet. That is not a failure: it is a starting point.

The real ROI is not what you think

A lot of the AI ROI conversation in our industry is about saving hours, reducing headcount, or unlocking revenue predictions. These things matter. But the most consistently undervalued return I see is much more straightforward: give your people a few minutes back each day.

That sounds modest, but it is not. Those minutes, accumulated across a team, across a season, mean more time with guests. More time developing skills. More time thinking rather than just doing. That is where the compounding value lives, and it is measurable if you choose to measure it.

The challenge is that most organisations build a business case with impressive numbers and then nobody actually tracks whether those numbers materialised. Private equity backed businesses tend to be far more disciplined about this, monitoring every penny of technology spend against outcomes. 

The rest of us need to take a page from that approach.

The talent and governance gap is real

One of the most candid moments in our panel was the acknowledgement that outside of the very largest hotel groups, most hospitality businesses simply do not have the in-house technical resource to design, deploy, and govern AI safely. This is not a criticism: it is just the reality of our industry.

The answer is not to wait. The answer is to seek the right guidance. Fractional expertise, specialist advisory, and technology partners who understand the operational context of hospitality are increasingly accessible. The mindset shift, particularly in smaller and independent organisations, is already happening.

From a governance perspective, the starting point for any organisation, regardless of size, is policy. Clear, documented rules about what AI tools employees can use, what data can and cannot enter those systems, and how documents should be classified. This is not glamorous work. But without it, connecting your SharePoint to an AI assistant, for example, can expose things you absolutely did not intend to expose. We ran that experiment internally. It was a useful lesson.

Every single one of your employees is already using AI in some form. The question is whether your organisation is in control of how.