It’s 2020, my name is Jack and I work at the front desk of a busy 3-star hotel at Amsterdam airport called the Grand Hotel (fictitious). Our clientele is the usual airport hotel clientele you would expect, crews, holidaymakers, business travelers, and in low season some tour groups. There are 300 rooms, and we have a large restaurant and a lobby bar as well as gym, meeting rooms etc. This morning was a particularly hard morning at checkout though. The car parking validator broke down and that just transcended us into a bad place. We are close to the airport, so we charge for parking, but some guests pay in advance and we must check this in the booking and validate the ticket with the machine, and if the machine is down then we must call the third-party vendor and so on and so forth. As it stands a normal check out can take around 4 to 5 minutes, so this pain of dealing with the machine hurt hard. At our desk the PDQ machine is not interfaced to PMS, so we have auth code messing around to deal with as well as frequent disagreements about final invoices that incur a login to the distribution portal. Despite it all, and the grunts and the pissed-off guests I did feel really bad when a little 5 or 6-year-old girl who just wanted to show me her new dolly but I couldn’t even think straight as I was trying to get my stuttering computer to load all 1000+ rates while I tried to book a room for a guest over the phone so I just snapped a ”ONE MINUTE MISS”… that one really makes me feel bad, I remember when I was a kid going to hotels and they were always so nice at the front desk, but they didn’t have all these computer programs to deal with, life was so much easier without computers…
My dream is ‘Jack in 2030’ comes true, because Jack in 2020 is real and in pain, and this pain is a real and present danger that will erode true hospitality. Its time technology is leveraged to enable and not hinder our heroes in the hotels, and embracing AI, particularly Machine Learning is the way to achieve this.
”Machine Learning is the tool to remove the drudgery from our lives”, A simple and reassuring definition of ML according to Cassie Kozyrkov, Chief Decision Scientist at Google. Unfortunately we the technologists have brought more drudgery to our operations teams, unintentionally, in an attempt to push boundaries and achieve the ever-elusive competitive, technological edge over our competitors, to the point that a receptionist often no longer looks up from their screens during a check-in/out. I spend more than half of my working life in the trenches extracting and eliciting business needs/requirements and user perceptions at the front desk, in housekeeping, in the kitchen etc. to know and understand the impact of our technologies on the end-user. Too often hotel management companies only hone in on one thing, the profit margins. Yes, there may be a forecasted 40 million more international departures in 2020 and plenty new stock coming online with nice shiny chandeliers, but a repeating guest is a happy and profitable guest. Why has AIRBNB taken so much market share so fast? Simple, it’s hospitable, it’s the rank me 5-star phenomenon. Some of the nicest and most hospitable ”hoteliers” I now know are AIRBNB hosts or traditional B&B owners. Unless you are paying top dollar the general perception of a 1/2/3* hotel guest expectation is on par with a low-cost airline. ”Hey, you get what you pay for”. Wrong! I started my career a long time ago at a 1* inn, in the west of Ireland, with metal keys and a paper book and those guests were the happiest I ever seen, why?, because I was human, my eyes were not on a screen they were on you, I knew what university John’s daughter was in, I knew Mary liked gravy on the side and I should never tell Joe the traveling sales man’s wife on the phone that he had a double room etc. I have met and talked with so many young, aspiring hoteliers over the last year across Europe, and they want to be engaged at the desk and talk with their guests and learn but the sheer mountain of tasks prevents them from doing so and so many are leaving the business as a result of this drudgery. Hospitality technology has not evolved in line with the generational trend and global technology trend. I visited a school this morning in Amsterdam and watched a class of 7-year old’s work diligently on their tablets on a task from the Digi board, but yet tonight their parents may check into a hotel and fill in a paper registration card with pen and paper, therein lies a significant gap.
What should hotel technology providers be focusing on for tomorrow? Guest expectations? Hoteliers expectations? Hospitality innovation? Neither, we should be catching up with global technological trends, trends that aid humanity and are already incepted in our daily lives. Machine Learning is right now analysing millions of mammograms across the planet, your bank is using Machine Learning to detect fraud, hey, if you drove to work this morning the traffic management was probably controlled on the highway by ML-enabled tools, or maybe it was just google maps powered by ai.google.
But WE have AI, says Vendor number 1, ok let’s just talk for a second about what AI is and how it can be interpreted in our industry. AI and Machine Learning, what’s the difference? Splunk.com has a nice simple e-book on this, ”5 Big Myths of AI and Machine Learning debunked”. Essentially, Machine Learning is a specific type of AI, that has access to carefully curated datasets and is trained on its decision process. Machinelearningmastery.com also has some easy to read articles on the learning aspect and indicate there are at least 14 types of learning in play to date. In respect to ML problem solving there are essentially 4 main models. Supervised learning: Where a user is tweaking and assisting the machine in its decisions (like a recommendation engine for a call centre agent in a hotel booking centre). Unsupervised learning: Where data is uncategorized (non-clustered) and the machine has to manipulate the data set itself and result (i.e.an automatic room assignment machine with data feeds from multiple sources). Reinforced learning: The machine now knows what it’s doing and works for reward (i.e. a revenue management system that has no human intervention at all, a ML process that is fed from everywhere and reacts accordingly. OYO’s revenue management algorithms are rumoured to be built with this vision in mind with 10’s of millions of updates per day already occurring). Deep Learning: Further down the line is deep learning for heightened rewarded learning in decision making (Check out Googles AlphaGo on this topic). The multitude of ML levels and AI advancements is evolving at pace, the commercialisation of Space amongst other things has brought a whole new bucket of money to the effort and the speed of advancement is impressive (i.e. SpaceX CIMON).
To automate or to teach? 40% of all technology start-ups that claimed to be AI-based or holding some form of AI (from all industries and formed in last 4 years in the EU) surveyed by a Venture Capital team in the UK in 2019 had no trace of AI technologies at all. So AI is a buzz word, and well the customers of those fake AI firms may still reap the reward of automation, I am not advocating anyone drop tool on automating things that should have long ago been automated and try to use ML for the purpose, automate what needs to be automated and inject ML where that is needed and not just for the sake of a marketing campaign, but we must all be careful in how we throw around these acronyms.
ML has no immediate ROI. Of course, we have been slow to gain any traction in our sector as like any other technology change (move to cloud etc), as such an initiative, does not show any immediate ROI nor likely to in any short term. ”How much will I make by the end of this year if I start AI now”, ”nothing, nada, zilch, but you will need to plough a lot of money into it for many years.” Sure, that’s likely to get a great round of applause at any behemoth hotel chain boardroom not to mention not getting in the door of a smaller hotel chain. Investment in technology continues to be on a drip-feed from hotelier executives and upsetting the apple cart a non-idea, the hope is disrupters like OYO, Airbnb etc fuelled by an enlarging disruptor technology offerings that harass and cajole legacy providers to follow suit in their strategies may in eventuality drag that investment of time and money to the fore.
Health tech started research in AI in the 1960’s today medical technology giants are many, like Philips Healthcare who exhibit a plethora of models and offerings, including for example the launch of an AI Platform in 2018 for general health tech to enable others to build on its foundations, but even after so many years of graft the advice is poignant “The quality of your AI is only as good as the quality of the data you feed into it,” according to Jeroen Tas, Chief Innovation & Strategy Officer Philips. However, the result is they now have ML helping doctors to reach more patients, diagnose faster, and cure more often. Frost & Sullivan research predicts that AI in Health Tech will save the health industry $150 Billion by 2025. Use cases are multiple, from ML supervised learning from the physicians’ interactions to elevate learn in diagnosis trends, analysis whereby machines can spot the tiniest anomaly at 4 am when a doctor’s eyes are weary, patient treatment plan defining and updating based on real-time symptoms analysis combined with regional information feeds to revenue cycle platforms for claims and payments. Medicalfuturist.com is a great source of inspiration in respect to health tech trends.
So, the problem is there, so how do we fix it?
- To technology companies: I look to the entire technology industry to reassess your 5-year plans and most importantly to the start-ups, to give more to the effort, to have break out days with ML engineers and before you add a story to Jira, ask yourself ”Could we achieve this with supervised ML?” or ”I wonder how health tech has addressed this in hospitals?”
- To the hoteliers: Create workgroups, assign budget, ML is now very accessible to all and either through companies like Trifork who can be commissioned to build for you or go on your own with your own digital factory where you can get up and running pretty quick with ai.google or Amazons do it yourself on ml.aws.
- To the hoteliers: Push your current providers for their AI/ML roadmaps and intentions, push your future vendors and align your technology strategy to include ML. Sacrifice a trip to Hitec for a trip to AI Summit Montreal this year.
- To everyone: It’s time to ringfence the brightest brains we have across our industry while learning and interacting with health tech/fintech/AI giants most experienced to learn, collaborate, research, define and build for 10 years from now and not just to end of year figures.
The reality is, Jack in 2030 is the same Jack, he still has the same job, but Machine Learning enriched technology built with his life, his user journey, his customers journey and expected experience in mind turned it around and made his guests happy and he with the same satisfied feeling I had all those years ago with my metal keys. In 2030 the guest will have paid everything or has a virtual wallet, the key is RFID, and booking a room is done in seconds aided by a hyper-personalized recommendation engine. Jack will still have a desk and a screen, for sure, he doesn’t like wandering around sofas asking people if they want to check in on a tablet, but his desktop screen will be intuitive, projected onto the desk itself perhaps and resulting in less drudgery and more time to crouch down for a minute to chat to a little girl about her wobbly tooth which ultimately is the reason he has stayed in this business all his life, and the reason Jessica might just follow in his footsteps.
Credit: Google News
