Introducing Sahra the sightseeing robot
June 3rd, 2020
The urban sightseeing industry was already on the precipice of significant change as a result of ongoing digitisation and then COVID-19 hit.
There have been so many predictions about the future of urban tourism, and the future of travel, I don’t really want to dwell on this too much. However three fundamental COVID-19 changes DO create the environment where Sahra, my AI powered tour guide that I introduce later in this article, can flourish.
In London, the BBC reports:
NEW: 20 passenger limit to be introduced on double decker buses in London & 6-10 limit on single deckers. #tflboard
— Tom Edwards (@BBCTomEdwards) June 2, 2020
To see what this does to tour buses, see how Zeelo (a bus mobility business) has calculated that physical distancing will impact their capacity.
Will we continue to have 2 metre physical distancing? Will it go to 1 metre? Who knows. But best guess is that tour buses are out of use for the foreseeable.
With international travel curtailed (or stopped completely), companies that are historically international tourism focussed have pivoted to domestic markets.
This is great but locals do not book the same sightseeing products as international tourists. For example they are less inclined to book a tour guide. They are much more inclined to be self-guided in some way.
Even if people DO travel, for the next few years people are going to be much more budget conscious. So many people have lost their jobs or otherwise taken on debt that leisure time budgets are going to be squeezed.
Some tour operators are lucky, they can operate physically distanced.
Some tour operators are lucky, they have a strong domestic tourism market.
Others are unlucky and are hit by both challenges.
We need new sightseeing product, in smaller vehicles, at low prices. Simple!
In our sector we have seen 3 significant technological advances in the last 20 years:
- Web – This created online travel agents & Software as a Service platforms such as web based reservation systems
- Social – This created reviews (e.g. TripAdvisor), Airbnb, P2P tours & activities
- Mobile – This created the need for online travel agents to connect to reservation systems as with the increase in last minute bookings you needed last minute availability and bookability
As an entrepreneur I have been trying to work out what the next in the list is. I thought it was going to be autonomous vehicles (such as self driving sightseeing vehicles) but now I believe it will be AI – artificial intelligence. Of course AI is a required part of autonomous vehicles so they are closely related as a concept.
What does AI give us?
AI enables computers to take on human characteristics. We can create digital humans that can deliver a similar sightseeing experience as human tour guides. What!?
This changes the industry.
Upto now we have focussed on digital retail, human delivery. This is now digital retail, digital delivery.
The audio-tour companies went here first. However their experiences are high audio production experiences and delivered the same each time. AI enables us, with the right data, to create a much more human delivered experience, personalised to the individual customer.
These new experiences are:
- Self-guided (or AI tour guide guided, if you want to say that)
- Low cost – there is no marginal cost for delivery once initial sightseeing route design has been created
- Any vehicle – pedestrian, bike, scooter, private car
- Able to be designed for any customer type – a sightseeing route is just “data” that sits in a platform, just like a song sits in Spotify.
When you go back to the three core implications of COVID-19 that I started the article with, it turns out that AI tour guides can assist with addressing all three.
“The tour guide makes the tour”
AI tour guides are not going to be suitable for every tour:
My hypothesis is that there are three styles of sightseeing tour:
- First person – e.g. a winemaker explaining about how they make wine
- Third person – e.g. a tour guide explaining about how winemakers make wine
- Navigation oriented – e.g. a 5 stop food tour where the tour guide is there to take you to the 5 stops, in the right order, without you getting lost
I think AI tour guides work best with navigation oriented sightseeing tours and can work OK with sightseeing experiences delivered in the third person. First person sightseeing experiences are safe (for now!).
Additionally, AI tour guides do not impact activities where the guide is there for reasons beyond storytelling.
For an even longer list of tasks a tour guide does (that may or may not be possible for an AI tour guide to do), refer to my article from last october about micromobility based sightseeing.
Will AI tour guides take over the sector?
Hard to say.
I would however argue yes, and for two key reasons:
Convenience / cost will trump quality
Booking travel online is a poorer experience than booking via a human travel agent, especially for leisure experiences where an experienced human travel agent can emotionally connect with the customer, helping the customer discover the perfect thing to do. However, after 2000, booking online became the default for mainstream travel bookings. Convenience trumped quality.
AI tour guides may be less good than human tour guides. (And certainly will be initially). However they are going to be cheap, convenient, start-anytime…… and this is likely to be just enough to make them the default mainstream method for most sightseeing experiences.
Ease of optimisation is in three groups (e.g. reacting to customer feedback and changing staying at a spot to 30 minutes rather than 45 as people were losing attention):
Hardest to optimise
- Retailer to local to operator, with a reservation system inbetween – too many moving parts to effect change quickly. And no, standards are not going to help, or at least, not help enough
Good optimisation capability
- In house tour operators – e.g. GetYourGuide Originals – they can optimise faster than retailer to local tour operator, one reason why GetYourGuide Originals will be the best human delivered model over time
- Digital platforms – AI tour guides, audio-tours, games (like Questo)
With digital tour delivery we can A/B test different route designs and work out what is best. YES existing tour operators do optimise over time, but digitally delivered experiences can iterate fast. Speed wins.
Who is building AI tour guides?
Many of the bigger tech players are building the base technology that could become a tour guide of the future.
Amazon has over 10,000 employees working on Alexa. They are known to have interest in hospitality and will do this sometime (or will enable independent skill developers to build this, even if Amazon do not do this work themselves)
Facebook were working on an AI tour guide in 2018, taking a vision approach. See this article. “The researchers at FAIR (Facebook AI Research) say they haven’t been able to create AI agents that can tackle this problem yet. (Why? “Because it’s super hard!”) But they expect teams to start building bots that can guide virtual tourists sort of competently in the next few years”
Google has Meena that could be applied to AI tour guides. “Compared to the SSA scores achieved by other chabots, our SSA score of 72% is not far from the 86% SSA achieved by the average person. The full version of Meena, which has a filtering mechanism and tuned decoding, further advances the SSA score to 79%.”
Apple has Siri.
This isn’t going to go away. These are your future general search engines that everyone optimises for today.
Sahra the sightseeing robot
To give you a better idea what an AI tour guide may look like, here is where we have got to so far with the AI tour guide I am building, Sahra. (Sightseeing Autonomous Hospitality Robot by Autoura).
Conversational user interface
Currently chat based, will be voice ultimately.
Consumer selects a sightseeing route…… and is then navigated turn by turn
What does it look like at the Autoura platform end?
We base our routes on true availability. e.g. if a route incorporates dinner with a local, once that space is booked up, that route is no longer available for booking.
In this example Borough Market (food market in London) is closed Sundays. Bermondsey Beer Mile (craft beer hotspot) is open Friday, Saturday & Sunday only….. hence for this sightseeing experience that incorporates visiting both locations, we know it is only available Friday & Saturday.
We are writing dialogue, vs training the dialogue (like other AI platforms). e.g. in this example, if the customer is going to a zoo, we may ask “Did you see the monkeys?”. Depending upon what the customer replies, we have our next line already scripted.
You could call this simulated intelligence rather than AI. In the end, it doesn’t matter if it is simulated intelligence or artificial intelligence – what is important here is the customer experience.
Ultimately Sahra will be embedded within autonomous sightseeing vehicles delivering a fully featured sightseeing experience. To those wondering, no, not given up on that yet. Anything but 😉
It is hard to predict anything right now, with COVID-19 here, but here are two to leave you with:
- The future urban sightseeing battle is NOT retail distribution, it is experience delivery. Customer experience, personalisation, cost, integration with citywide systems…. will all be key. I predict the battle will be between vertically integrated humans (e.g. GetYourGuide Originals & perhaps WithLocals) and digital delivery platforms
- P2P experiences failed over the last decade because they had to be delivered by amateur tour guides. With this new tech, AI tour guides do the delivery, enabling a passionate local to focus on experience design. This will see a resurgence of the concept……. what do we call it – P2B with B for bot?
Final thought – if you want to be first with an AI tour guide in your destination, shout! Signing up tour operators now…….
Image: Flickr: JMorgan
This content is protected by copyright. Link sharing is encouraged but duplication and redistribution is illegal