How decentralisation & AI will dominate the next decade and beyond in sightseeing

Too much AI debate in the sightseeing sector relates to generative AI such as text, image & video generation. This article outlines why we should prioritise addressing other forms of AI first, and how we can incorporate decentralised technology approaches within an overall AI strategy.

First, why is everyone distracted by generative AI? I believe this is because for the flight & hotel sector generative AI is the (retail) AI that matters. For the sightseeing sector it is further down the priority list.

Multi-sector travel trade press, when covering AI, will focus on generative AI as that is what resonates for their hotel & airline sector readership. They are right to do so. But all this discussion distracts us in from what is important in sightseeing.

Key differences between flight/hotel AI & sightseeing AI

With a flight, the product sold is a seat to get from A>B. With a hotel, the product sold is a bed to sleep in from 11pm to 7am.

There are additional layers to this core product mostly focussed on how to make your bed or seat more attractive than someone else’s. This comes down to brand, to distribution, to economies of scale, to service. In the mix with other solutions, generative AI helps companies differentiate in that highly competitive (& lucrative) market.

In the sightseeing sector we will primarily use AI for designing & operating new experiences, not retailing existing experiences. The sightseeing product itself changes as a result of AI. Once you change the product, this changes how you distribute and even what companies & platforms you need between the consumer and their service.

i.e. within flights & hotels, with AI, the product remains constant (it is still a seat or a bed), in sightseeing the product changes completely.

It is this upcoming AI lead product change that we should spend more time debating in sightseeing circles, rather than generative AI. Once the product changes are settled, then we can revert to talking about retail impacts & opportunities.

Using AI to operate vehicle tours

Lets just take one example, vehicle based tours in urban locations. This sub-sector is interesting as the sector is dominated by larger tour operator companies, it is capital intensive to buy tour bus vehicles, and these operators often operate quite a mainstream high volume product. Exactly the kind of business model that AI disrupts.

Who are these disruptors? They are out there today, just not scaled up yet.

For example there is a US/Chinese travel industry company that operates vehicle experiences in multiple US cities solely using AI. They currently do 10,000 trips a week (vehicle not pax, I believe) including consumer airport transfers and have recently announced they will scale up to 100,000 trips a week by US summer 2024.

Think about that a second, 100,000 local experiences a week using AI, no humans. And thats just the start.

Who is this company you might ask? Well it is Google and their autonomous vehicle division Waymo. You may have heard of Google. (The Chinese part is Geely, their vehicle partner)

They are not alone, their peer competitors are Microsoft (via Cruise) & Amazon (via Zoox).

Likely game over for tour bus operators in urban locations in the USA when these companies add a tourism layer to their mobility platforms.

This video (and apologies if you have seen it before) is my best explanation as to what the product impact is…. pay attention to the shift to long-tail experiences as a result of AI. Sure, a poetry tour in San Francisco is not a mainstream experience, that is the point.

Decentralisation

However, don’t lose all hope yet, there are some interesting technologies coming down the track that could swing the balance back to local tour operators, helping neutralise the disruptive forces that will come from new entrants using AI to operate experiences at scale.

One key approach is decentralisation (and the personalisation that it enables).

The two groups pushing this approach forward are:

• Decentralized Identity Foundation (DIF) – “Join us in developing the foundational components of an open, standards-based, decentralized identity ecosystem for people, organizations, apps, and devices”
• Trust Over IP – “We’re an independent project hosted at the Linux Foundation, working with pan-industry support from leading organizations around the world. Our mission is to provide a robust, common standard and complete architecture for Internet-scale digital trust.”

Within DIF, I am member of the the Hospitality & Travel SIG (Special Interest Group). This week we have published our initial concept for decentralised travel profiles. You can read about it here. (See my name on the collaborators list for that standard)

Within Trust Over IP, I am a member of the Attraction Pass Task Force. That group is working on the governance and technology aspects required to decentralise city wide attraction passes.

These are open groups and if you want to join, join.

Why is decentralisation important to sightseeing?

In sightseeing our customers are mostly single transaction. A customer comes to our destination once, and our tour once. Even if the customer comes back to the same destination at a future point they are unlikely to try the same tour or even the same tour operator.

Evidence of this is the lack of loyalty schemes in the sightseeing sector. If you go to an airline or hotel conference, a dominant topic is loyalty. If you go to a sightseeing conference loyalty rarely gets a mention.

As a result of all these single transaction customers there is no motivation to store travel preferences “for the next time” as there generally is no next time. If a tour operator learns that a guest is vegetarian, there is nothing to be usefully done with that information.

With decentralisation we store these preferences in a digital travel wallet. This wallet can be on someone’s mobile device, or it can be cloud hosted, or both. The key is that the consumer can share their preferences with any entity at any point. The consumer is in control and only they can share their own preference data.

So if the consumer wants to share with the food tour company that they are vegetarian, they can. If they want to share with the taxi company that they are vegetarian, they can. (Helpful if you want to chat to the taxi AI and ask where can we get some food now?). Not all data fields need to be shared with each interaction, just those fields required for that specific interaction.

This is a “configure once, use many times” model so people will make the effort to configure their preferences and keep them updated.

As in sightseeing we have no long term preferences storage, this decentralised travel profile will be a significant industry structure improvement.

How will this help with AI?

The AI that everyone wants to talk about (and with) is ChatGPT. But ChatGPT is centrally managed and with plugins it is a centralising force.

The future though is AI in everything, each LLM (Large Language Model) trained for their own purpose. e.g. when you are in a taxi you are talking to the taxi AI. When you are on a tour, you are talking to the tour guide AI, when you are in a visitor centre, you are talking to the visitor centre AI.

What decentralisation of the travel profile does is enable each of these AIs to have access to the same consumer preferences. It enables any local tour business to build tech that can access these preferences and offer amazing personalised experiences (digital & non-digital). It combats centralisation as it will not be the OTA (Online Travel Agent) profile that the local tour operator has to get permission to access, but an open system, open to any innovator to use.

Thoughts

Early movers of this decentralised approach within sightseeing should be the global franchise / membership groups as they have often struggled with creating single customer profiles over multiple member entities for legal reasons relating to personal data transfer. Decentralised travel profiles fixes this.

Decentralised preference data and AI lead autonomous vehicles are inevitable. The technologists at rez tech companies understand this, but they are locked into the travel agent > rez tech > tour operator industry model. If they are looking for a good wave to surf, they missed the autonomous vehicle one, but the decentralised wave is looking right sized for the next generation of travel tech startups to surf….

Time to build.

Image: Dall-E