HTE 2026: plenty of AI, hardly any guest data

The Hospitality Tech Expo filled ExCeL London on 29 and 30 September. Once it was over, we read what the industry had posted about it: 696 public LinkedIn posts, written between 14 September and 1 October.

One number stood out. AI came up in 114 posts. Guest data came up in 13.

What did HTE 2026 talk about?

A post can count towards more than one topic. Method at the end.

Payments leading is no surprise. Every venue takes money, and a payments decision touches every shift. AI in second place is no surprise either. It is the question every operator is being asked about, and every vendor is being asked to answer.

The surprise is how rarely the two met the thing they depend on. Of the 114 posts about AI, 4 mentioned guest data or a CRM.

Is AI worth it without clean guest data?

An AI tool for a venue is only as good as what it knows about the venue's guests. Ask it who to invite back for a quiet Tuesday and it needs to know who visited, when, what they spent and whether they said yes to marketing. Ask it to reply to a review and it does better knowing the reviewer is a regular on their fifteenth visit.

In most venues that information exists, but in pieces. The booking is in one system, the bill in another, the Wi-Fi login in a third and the review on Google. As one operator put it to us: "I've got everything everywhere, quite literally." Put AI on top of that and it is working from a fraction of the picture.

AI on scattered data makes a good demo and a poor tool. On the stand it answers every question. Back at the venue it is guessing, because the booking, the bill and the review it needs are sitting in three systems that do not know they belong to the same guest. At best that is a smaller result than you were promised. At worst it is smoke and mirrors: confident answers built on a fraction of the picture.

At HTE the industry mostly had the AI conversation without the guest data one.

What does AI on clean data look like?

It starts with the record, not the AI. At Stampede every booking, payment, Wi-Fi login and review lands on one guest profile as it happens, and our AI features work from that profile rather than sitting beside it as a separate tool. That is the difference between an AI that has to be told who a guest is and one that already knows.

Whoever you buy from, that is the test: does the AI read from one joined-up guest record, or from whichever system it happens to be plugged into?

Five questions to ask the vendors you met at HTE

If you came home with a bag of leaflets, these questions tell you whether a tool, AI or not, will add to your guest picture or split it further.

1. Where does the guest data from your product end up? If the answer is "in our system", ask how it reaches yours.

2. Will it recognise the same guest across bookings, Wi-Fi, payments and reviews? Five records for one person is five guesses.

3. Does it sync both ways with my till, or only one way? One-way means spend arrives late or not at all.

4. Who owns the guest list if I leave? Ask for the export format before you sign, not after.

5. How does the price change as I grow? Per cover, per contact and per venue pricing behave very differently on a busy December.

How we did this

We collected public LinkedIn posts that mentioned the show between 14 September and 1 October 2026, using the official hashtags, the show's name and the organiser's and exhibitors' pages. We removed duplicates and checked each post was about HTE rather than one of the shows sharing the venue, or an unrelated event with a similar name. 696 posts remained.

Topics are counted by keyword in the post text, so one post can count towards several. "Guest data" means the post named guest or customer data, first-party data, a CRM or guest profiles; a passing mention of "data" alone does not count. LinkedIn search returns a sample, not every post, so read the figures as the shape of the conversation rather than a census. We did not name, contact or target anyone in the data.

Were you at HTE?

If the guest record is the conversation you came to have, book a thirty-minute demo on your own venue, or start a 30-day free trial.