|
Insight Comment |
Artificial intelligence (AI) is an umbrella term which refers to a variety of intelligent programs. Every time we run a search on Google, AI is working behind the scenes. Word and sentence predictions on Gmail or Whatsapp are powered by AI and every advert, notification and recommendation that we receive on social media is the result of AI.
Tristan Gadsby, CEO and co-founder, Alliants, commented: “AI is everywhere in our daily lives but it’s still common for people to have misconceptions about it. AI is not a hyper-specialised branch of computer science, nor is it anything particularly new. This is the technology that beat world chess champion Garry Kasparov back in 1996. It’s the technology that allows you to do online translations. There are plenty of ways AI can enhance the guest experience and streamline and simplify hotel operations, but the crucial first step is for owners and operators to identify where the use of technology can drive business value.”
The hotel industry has been using AI via revenue management systems and demand forecasting for decades, but this article will explore other areas of hotel operations where AI is used. To do this, we shall follow the guest journey, from research and booking to post-visit communications.
Machine learning (ML) is a subset of AI that refers to programmes that modify themselves without human intervention. One of the most common uses of ML is to make product recommendations.
With vast inventories for customers to choose from, the OTA duopoly (Booking.com and Expedia) and Airbnb are widely regarded to excel in this capacity.
Around two thirds of travellers visit an OTA to gather information first before booking directly with their chosen hotel. Hotels want their websites and booking engines to be efficient and reduce the risk that users will return to the OTA to complete their transactions.
Ireland-based Dalata Hotel Group recently integrated their website and booking engine with an AI platform provided by another Irish company, Arvoia. The platform ‘learns’ about the needs of each guest from their movements on the website and consequently alters the booking engine and website to prioritise the most relevant rooms, hotel amenities and content rather than just showing room rates in price order.
Dalata Hotel Group saw average booking value per guest and conversion rates rise consistently over the first three months of deployment.
How much do hoteliers know about their guests before they arrive? How likely are they to complain or be so satisfied that they will endorse the hotel on a review site?
Jacques-Olivier Chauvin, CEO, Fauchon Hospitality, developed an algorithm to specifically address these questions. The algorithm is used at the group’s hotels in Paris and Kyoto as well as others outside the Fauchon brand.
How does it work? Chauvin explained: “On a daily basis the list of clients arriving the following morning is fed into the algorithm and their data is fully anonymised so we just get the results of 15 variables per client describing their nationality, demographics, the characteristics of their stay, duration, category of room, reason for coming, etc.”
“Each client is then mapped to roughly 800 profile clusters. For each of these clusters there is a probability that the client will be super satisfied enough to share on social media.”
Every morning, the reception team receives a report establishing on a scale from one to five stars the probability of satisfaction for each single guest on arrival. One star means a high probability that the guest will complain and five stars means that they are likely to leave a positive review on social media.
The algorithm has found that an American couple over the age of 60 in Paris for a long stay are likely to be super-satisfied, while a Korean businessman travelling alone is likely to complain. Chauvin is convinced that the algorithm has helped Fauchon L’Hôtel Paris achieve consistently high rankings on TripAdvisor since opening in 2018.
“I wouldn’t say it’s 100% reliable, but it focuses the attention of the staff onto particular people which is very good,” he commented.
During their stay, if you have permission to communicate with guests via their smartphones, a messaging platform utilises AI in several ways, including instant translations, automated replies (chatbot), suggested responses (whisperbot) and sending conversations to the right departments.
Such a platform can also create personalised digital itineraries and recommendations for guests which are likely to result in higher ancillary spend.
Sea Containers London, part of the Lore Group, recently chose a platform provided by Alliants that allows guests to communicate with staff on the messaging app that they prefer, whether that is Whatsapp, Facebook Messenger or SMS. The platform gives all employees a real-time view of guest requests and itineraries.
Paul Rasche, director of IT, Sea Containers London, said: “The simplicity of the solution has meant uptake has been very high, with less time for our staff spent answering calls and more requests turned around a lot faster.”
At the Blackstone-owned Cosmopolitan in Las Vegas, the chatbot has a sassy personality. On arrival, guests receive the number for Rose, the digital concierge, who advertises her services as: “Resident mischief maker. If you’re looking for trouble, I can hook you up with the best we have to offer.”
Once the guest has left, AI can help with automated post-visit communications. Typically these include a message soliciting feedback and/or invitations to review the hotel. If the guest has opted-in, post-stay targeted marketing can be powered by an algorithm based on customer segments and individual data such as birthdays.
In order to manage online reviews, hotels can use reputation management software which costs $50-$80 per month per property. It finds and gather all mentions of the hotel from disparate online sources and can reply to reviews automatically.
Given that a significant percentage of online reviews are fake, or generated artificially, we have a situation here where a machine is writing a review and another machine is replying.
In other use cases, Aimbridge Hospitality said that it is trialling AI in some of its accounting procedures – removing the people element from repetitive processes - with very positive results so far.
Machine learning technology can also reduce food waste in commercial kitchens. Steven Smalley, a Hilton executive chef in Dubai, said: “We’ve seen a 70% reduction of waste which is for the chefs’ benefit because they are not wasting their time overproducing food.”
When considering an investment in AI technology, it is essential for hoteliers to establish use cases first, what AI will deliver, and what the ROI will be, said consultant Peter Russell of Russell Partnership Collection: “It might be that it costs a lot to bring in but it might save huge amounts.”
Upgrading customer relationship management (CRM) systems and/or moving towards Amazon-style personalisation is potentially the most complex and expensive application of AI, he said. Although the ROI might be significant, calculating it is not straightforward.
Russell continued: “Or an AI solution might be quite low cost but not really save anything , in which case, why are we doing it?”
How many hours of staff time might reputation management software save? For example, replying to reviews manually costs $80 per month in staff labour and so does the software, which cuts three-quarters of labour costs, because you don’t fully trust the software to send replies without checking them first, so the AI solution would still end up costing more at $100 a month.
With guest messaging platforms, the use case is more clearly defined, said Russell: “You know that guests will have certain questions. You are removing a burden on staff and providing an elevated customer experience. It’s one of the big areas where I see more instantaneous results.”