By Alasdair Snow, Co-founder at Triptease
There's a new wave of automation coming, and we believe that it will bring first-class luxury experiences within reach of most hoteliers and guests for the very first time.
Automation has been an obvious tactic for us, both because of the benefits it can bring to the guest but also because of the weight it can lift from the shoulders of already over-encumbered hoteliers. So let's talk about what automation means.
Automation is nothing new
A lot of the time when we think about automation we think about robots, like Boston Dynamics' SpotMini robot dog. It's undeniably a little unnerving to see a robot dog turning a handle and opening the door. These machines have learned to navigate their environment. They've worked out how to do it themselves - and for me, the really scary thing is that they're working as a team! There's a huge fear of the unknown that arrives whenever we see something like this. Where will it lead us?
Let's look back at automation through history before we get too caught up in our existential dread.
Automation is nothing new, and it's definitely not all about robots. Indeed, automation is arguably what makes us fundamentally human. Technical progress has been nothing if not a steady march of automating increasingly complex repetitive tasks. Rather than ripping our hands to shreds trying to do everything ourselves, we have always worked to delegate that burden to technology. As humans we're constantly looking to drive efficiency to free up our time to do other things. Automation is the drumbeat of human progress.
But, just like the knee-jerk jolt of unease most of us get when we watch those clever robot dogs opening a door, there has always been a fear of automation. Think of the Luddites. In case you haven't been brushing up on your history recently, the Luddites were highly-skilled artisanal cloth weavers in the early 19th century. They were particularly upset about the mechanization of the loom and how it enabled unskilled laborers to manufacture cloth. So, they smashed the looms. As a result, 'Luddite' has become a byword for people who aren't too happy about technological advancement. In the 21st century, their place has been taken in part by headline writers:
The fears on display here - of losing jobs, of being ruled by the machines - echo those felt by many at the advent of the Industrial Revolution. The rate of automation indeed exploded in the 19th century, but it didn't destroy economic employment. It raised the level of economic activity, creating more wealth, more demand, and more jobs. New technology does not lead to higher overall employment in the economy. Rather than destroying jobs, it changes the composition of the economy itself. In fact, 85% of jobs the next generation will do haven't been invented yet.
This new era is driven by exponential technical growth
So, automation is nothing new - but we're definitely entering a new era. In 2017 Ke Jie of China, the number one Go player in the world, was vanquished 3-0 by AlphaGo Master, developed by Google Deepmind. Go is an ancient abstract strategy board game, thought of by many as the most complex in the world. It has been around for millennia, but AlphaGo Master was deploying tactics and strategies that Ke Jie couldn't have dreamed of. Ke Jie's post-match comments after the match highlight the extent to which AlphaGo Master was operating on a different plane:
"To me it's like a god of Go. A god that can crush all that defies it [...] AlphaGo can see the whole universe of Go. I can only see a small area around me."
This really speaks to the new agenda of automation driven by exponential technical growth. We have taught machines to do repetitive processes on our behalf - check in guests, check disparities. Now we can teach machines to identify repetitive processes and create novel solutions to them all by themselves.
This is machine learning. Rather than teaching a machine how to do something, we give it a problem and it works out how to do it by itself. Above you can see Google Deepmind's AI teaching itself to walk. It's been given a basic anatomy, limbs, the ability to move and set an objective, get from point A to point B. From that, it's worked out how to walk, jump balance, handle gravity... it doesn't look much like you or I, but it gets the job done.
What's enabled this is massive exponential growth in computing power? Intel's co-founder Gordon Moore observed just over 50 years ago that the number of components in an integrated circuit doubled every year, and projected the same pattern to continue - and he was right. The growth of computing power has not been linear but an extraordinary exponential curve.
It's very important to understand exponential growth. It moves very fast, and shoots up in a way that can feel sudden and unexpected. You might have heard that if you take 30 steps, you'll travel around 30 metres. But if you take 30 exponential steps, you'll go to the moon.
We can see this pace of growth across lots of areas: the ability for machines to recognise images, understand speech, and perform natural language processing is advancing at a phenomenal rate. Most importantly, the companies that are advancing these abilities are sharing their technology and research - Google with their Vision API, IBM with Watson. They've all learned that there's lots more to be gained by making this stuff publicly available than by keeping it behind closed doors.
In previous rounds of automation, the machines replaced our muscle power. This time they're replacing our cognitive abilities - and are doing so in a way that we can't even come close to competing with.
Luxury experiences will be within reach of everyone
I started by saying that this new phase of automation would bring luxury experiences within the reach of almost all hoteliers and guests. I want to talk about some specific applications of this form of automation that could have an enormous impact on the quality of guest experience.
What distinguishes a 5-star hotel from a 3-star? A great deal of it is to do with the luxury of a meaningful relationship between hotelier and guest, where the hotelier knows who the guest is and what they like, and can tailor their experience to them personally.
Automation can help us do that in a totally new way. Machine learning can be applied to innocuous data points and turn them into rich, meaningful guest profiles, giving an in-depth insight far greater than could ever be achieved at the check-in desk. Just a couple of pieces of information - time of booking, search parameters, location - would be enough for a trained machine to compare one guest against the thousands of others it has learned from, then construct a fully fleshed-out guest profile from your hitherto anonymous online booking. Then, it's up to you what you do with it. It could be used to optimize marketing automation tools, or hotel staff could use it to enrich the in-stay experience.
Knowing your guests in detail means you can start to tailor their experience perfectly to their needs. At Triptease, we help hotels think about how to send relevant and attentive content to guests to help them navigate their site and make a direct booking. Something I've personally spent time thinking about is how websites might adapt their content to convey the most relevant information to each individual guest. So, rather than having three preset content 'buckets' that you put guests in, you have content that is dynamically reformulating according to the needs and characteristics of each specific visitor. Generative neural networks could be used to create this personalised content - they already are elsewhere. Wordsmith by Automated Insights is being used by the Associated Press to write quarterly earnings reports, business news articles and sports reports.
Luxury is about tailoring the experience, but it's also about listening attentively and giving each guest one-on-one attention. With Natural Language Processing, this is possible. Machines can parse meaning from human speech and formulate an appropriate response. A digital assistant at your hotel could talk to all of your guests simultaneously: the luxury experience at scale.
A key part of the Triptease platform is our live chat, Front Desk. The purpose of Front Desk is to extend the hospitality experience to people browsing your website. It's about bringing hospitality online. We know that a lot of potential guests run into trouble or have questions before making a booking. Front Desk allows the hotel to field those enquiries and facilitate a direct booking.
We're now bringing our learnings from Front Desk together with a lot of the technology I've discussed above. Introducing: Auto-Agent.
Auto-Agent analyzes guest requests and sends informative replies. Through automation we can deliver the kind of exceptional online service that customers expect but can only be achieved with significant cost of training and staff. So, it's great for the hotel. Machines don't sleep. They don't get tired. They don't get cranky. They don't become irrational. They don't need a pay rise.
It's also great for the guests. As consumers, we've grown accustomed to always-on, real-time messaging experiences, but most hotels are still struggling to cater to that demand. We know first-hand how important getting this right is to hoteliers. The story is there in the numbers: we see a big initial jump in online conversion rate when hotels begin to manage chats via Front Desk. But the jump is bigger still when the bot works with the human to handle all enquiries. Why? Well, it gets back to the guest straight away. Politeness is programmed in. And, ultimately, we're teaching it to understand what questions to ask and how to steer the guest towards making a direct booking.
Cognitive automation frees up staff for richer guest engagement
I'm not preaching the advent of general artificial intelligence, or that the machines will put us all out of our jobs. The near-term consequence of this sort of automation is that staff will be freed to develop even deeper relationships with guests.
At this point, it's worth noting Moravec's paradox. Hans Moravec was part of a team of AI researchers in the 1980s whose essential conclusion was, contrary to traditional assumptions, 'hard' tasks require little computation whilst 'easy' tasks require an enormous amount. Moravec wrote: "it is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility."
What does this mean? Machines are limited. Intelligence is very narrow, and handling sensory inputs is really hard. AlphaGo Master can play Go better than you, but it couldn't help a child tie their shoelaces. A bot can easily field questions about your breakfast menu, but it couldn't judge the respective moods and needs of a family arriving at a hotel.
It's still very difficult for computers to navigate this messy and complex world - but they can navigate certain tasks far better than a human ever could. Let them do the hard stuff and use the time you reclaim to attend to more complex guest needs. Those Associated Press journalists having their articles written by Wordsmith aren't out of work - they're just no longer churning out corporate earnings stories, and have moved onto meatier big-picture articles. The reception staff at CitizenM aren't job hunting as a result of automated check-in desks; they now make coffees for people arriving and chat to them about the local area.
The next wave of automation will deliver incredible guest experiences at very low cost. This is the era of bringing hospitality online.