Can Independent Hotels Afford to Wait for AI?

I remember listening to Barry Sternlicht talk about the early days of online travel agencies and wondering how the big hotel companies could have missed the opportunity.

These were the industry's heavyweights. They had the inventory, the brands, the customers and the expertise. Yet much of the online distribution business that followed was built by companies outside the hotel industry.

I can understand how it happened. Hotels were comfortable with the business they knew. We had spent decades building relationships with wholesalers, travel agents and corporate accounts. I remember hearing newcomers dismissed as small players. Another travel agency. What difference could they possibly make?

We know the answer now.

To be fair, the hotel companies weren't entirely asleep. There were attempts to build a common online booking platform, including Travelweb. And online travel agencies brought genuine benefits. They introduced hotels to new customers, opened markets and gave independent properties visibility they might never have achieved on their own.

But they also changed who held the power in hotel distribution.

I used to believe direct bookings were always significantly better for the hotel. Then we started examining distribution costs more closely across the group. Direct bookings were generally cheaper than OTAs, but the difference depended on which costs we included. Sales teams, reservations, marketing and loyalty programmes all cost money, even when those expenses don't appear against an individual booking. It changed how I looked at distribution. The question wasn't simply where the booking came from, but what the hotel actually earned from it.

Around the same time, I was noticing something about loyalty programmes. Guests increasingly belonged to several major hotel programmes, much as people carried credit cards from competing banks. They could collect points almost wherever they stayed. Membership was no longer a particularly reliable indication of preference.

Even today, when I look at a hotel's commercial performance, one of my first questions is what percentage of room revenue comes through OTAs and the hotel's own website. I compare that with what I know of the market. If the numbers look weak, I check whether the hotel is visible online, whether its rates are competitive and whether the relevant channels are actually open and bookable.

Sometimes management simply doesn't know how much business it may be missing.

When revenue falls, the instinct is often to send the sales team out for more calls and more visits to the same traditional accounts. Those relationships still matter. But there comes a point when you have to ask whether you're beating a dying horse instead of finding out where the customers have gone.

Which brings me to AI.

Booking Holdings recently announced Vayo, an AI-powered personal travel agent, initially in private beta. The idea is that travellers could use an agent to plan a trip and find accommodation, potentially completing bookings through different channels.

It is early days, and I have no idea whether Vayo will succeed. What interests me is what happens if travellers increasingly turn to an AI agent before they turn to a hotel, an OTA or even a search engine.

The hotel may still receive the reservation and deliver the experience. But the business helping the traveller decide where to stay could develop considerable influence over which properties are considered in the first place.

We have seen something like this before. The difference is that the AI market has not yet settled.

The established online travel companies have considerable advantages. So do Google and the other technology giants. But smaller companies are experimenting with different approaches, and some may find opportunities the incumbents have overlooked.

The Davids of yesterday have become the Goliaths of today. There is still room for another David.

That could be particularly interesting for independent properties. A young technology company needs more than a clever product. It needs access to real customers, operating environments, feedback and evidence that its technology delivers commercial value.

Hotels can provide those things.

If I were advising an independent owner, I would look for opportunities to run a few controlled proofs of concept, perhaps over three to six months, ideally with the technology provider absorbing much of the development cost.

I would not start by buying expensive systems or committing to long contracts. I would look for promising companies that need hotel partners and see whether we could learn something useful together.

A group of independent hotels might even be able to offer a more attractive testing environment than one property acting alone. Singapore's tourism innovation programmes offer a useful example of technology businesses working with hospitality operators to test new ideas in real settings.

There are several areas worth exploring. One is helping travellers discover a hotel they might otherwise miss. Another is improving conversion once a potential guest shows interest. A third is using guest information more intelligently to encourage return visits.

Each needs a different commercial test.

If a new channel produces reservations, I want to know how much business is genuinely incremental and what contribution remains after acquisition and servicing costs. If a system promises savings, I want to understand the cost of integration, training and ongoing operation. If it claims to improve loyalty, I want to see whether guests actually come back.

I would also be careful about who controls the guest information, what happens when the pilot ends and whether the hotel can walk away without disrupting its business.

My cost-per-channel experience taught me not to accept attractive numbers until I understand what is included in them.

Of course, an owner may have perfectly good reasons to wait. A hotel with urgent maintenance needs, weak cash flow or operational problems has more immediate priorities. Even a supposedly free pilot consumes management time and can create problems if poorly managed.

And being first doesn't guarantee anything. A technology advantage can disappear quickly once competitors adopt the same tools.

Still, waiting has a cost of its own.

A promising startup may be more willing to listen to an independent hotel today than it will be after its product is established. Early partners may have opportunities to influence development or negotiate terms that won't be available later.

More importantly, a hotel that experiments carefully begins to understand what works before it has to make a much larger decision.

None of this changes what makes a good hotel. Location, personal service and an experience worth remembering still matter. AI may make hotels easier to compare, but it could also help a distinctive independent property reach travellers who would never otherwise have found it.

I don't think hotel owners need to predict which AI company will win. We didn't need perfect foresight to understand the economics of online distribution, either. We needed to look more carefully at what was happening and question some comfortable assumptions.

This time, we have an opportunity to get involved before someone else decides how the business will work.

I'd rather learn a little now, while we still have some influence, than discover a few years from today that we've left another important decision to somebody else.