October 2017, Las Vegas (USA)
How ‘agnostic’ Microsoft is marching into a brave new world
The applications for artificial intelligence (AI) in travel are moving beyond hype, but can it fundamentally disrupt the search and distribution landscape? Pamela Whitby has been finding out
“There is no doubt that the hype exceeds the reality, but the reality is still profound,” says Stuart Greif, a senior executive at Microsoft who has his eye closely on how everything from AI voice-enabled search to augmented reality, facial recognition, chat bots and more will impact travel and hospitality. While he believes we may be some way away from the dystopian reality that the naysayers are predicting about the rise of artificial intelligence, Greif argues that in travel “the potential exists for a complete disruption of the current search and distribution paradigm”.
So how is Microsoft positioning itself in this new world order?
Well as you would expect from a tech giant, it is looking to set global standards in the advancement of new technologies, and is doing so with some success. Microsoft produces more AI research than Google, Facebook and Apple, holds the most AI patents by a significant margin, and its image recognition technology (which Uber uses to validate its drivers) has won awards for outperforming the systems of firms like Tencent and Qualcomm.
Microsoft has also been active in the chatbot space. It created a Facebook Messenger chatbot for the The Venetian, the first hotel in Las Vegas to do so and a partnership with Skyscanner, involving a so-called flight search skill, allows “users to ask a range of handy, direct questions through Microsoft’s personal assistant”.
Aside from its strides in image recognition to create efficiencies, enable safety or simply delight customers, Microsoft is also leading the way in voice recognition. About a year and a half ago, explains Greif, “we achieved parity with human conversation speech, and on August 20th this year, we actually achieved a super human rate of cognitive speech recognition.”
There are, however, still limitations for certain sectors like air travel where safety is paramount. As Greif explains: “You need a level of accuracy that is approaching 100%, and over a few years, to have the confidence that a command won’t be misinterpreted if a pilot doesn’t look down”.
Making AI available to all
While Microsoft will continue to push the leading edge in AI, it’s not trying to be all things to all businesses and people – in the travel industry or, for that matter, elsewhere. This ‘agnostic’ approach is being driven from the top by CEO Satya Nadella who, according to Greif, has “digitally and culturally transformed the company” by striving to “be a learn-it-all not a know-it-all”.
So, while some developments remain under wraps, and may only be used for proprietary purposes because they are new and emerging, Microsoft wants to make AI available to everybody. For this reason, explains Greif, “we have taken an open approach to partnering with world class players where we see the opportunity to advancing our goals of helping every organisation and individual achieve more.”
In most cases, says Greif, “that means the partner is on Microsoft’s cloud, but we want to be where customers want to be and sometimes that even means partnering with competitors". The recent alliance with Amazon, to link their two voice-activated digital assistants – Alexa and Cortana is one example. In others cases, it could be that Microsoft develops for competitors’ platforms first such as apps for Android (Google) and IOS (Apple).
Microsoft is also looking to partner with disruptive startups such as the with marketing data platform, Amperity, which came out of stealth mode last week.
What Microsoft seems to have recognised is that with the way technology is moving, it’s becoming harder than ever for a single platform to dominate.
The rise of the ultimate meta
But what happens when voice-enabled search, powered by artificial intelligence can deliver a smaller number of more personalised recommendations that more accurately fit a customer’s parameters. Take this scenario: A Londoner travelling to Shanghai wants four to five star accommodation, close to a given address within a certain price range. He also wants the search to return his two favourite brands, even if they aren’t close to the location. The question then is who owns that search? Is that a digital assistant like Alexa and Cortana working together? Is it Siri, Bixby, or Facebook? Is it the operating system of the actual phone - Android or IOS? Or a metasearch, Google or brand.com?
According to Greif, it is against this backdrop that providers or intermediaries must ask: “Where do you stand in a world, where not just voice-enablement but the AI behind it, including bots as one manifestation of that, no longer follows typical search patterns. And what if AI can search all of those sites, possibly circumventing Google’s ad model, to become the ultimate meta if you will?”
In the interim what is likely is a wild west where incumbents find strategic ways to tie up with multiple parties
The answers to these questions, and how long this might take, remains unknown, but in the interim what is likely is a wild west where incumbents find strategic ways to tie up with multiple parties.
Whether it’s Kayak partnering with Alexa to enable users to search flights, hotels, and rental cars or Avis launching voice-powered car rental reservations, this is already happening today. But it remains limited to search; artificial intelligence capability still falls far short of actual transactions. Instead, it is being used behind the scenes in a way that users don’t notice. For example, AI bots can be used to answer a basic subset of questions before finally kicking up to human agent status. But they can also make recommendations and are becoming smarter and better by using both supervised (human intervention), and unsupervised (machine learning) techniques; just as Google search and Amazon recommendations, which weren’t so great at first, improved over time!
It’s early days, with most travel companies still going for the low-hanging fruit. “The near-term promise is focused on more discreet areas and use cases,” admits Greif, “because it’s still incredibly complicated to book travel”.
The scope of the Avis service, for example, was very narrowly defined within three parameters: i.) only for customers who had downloaded the app, ii.) they were able to look at making a booking; and iii.) receive confirmation of those details.
Early days it may be, but that doesn’t mean travel brands should be standing still. As Bill Gates once said: “Most people over estimate what they can do in one year, and underestimate what they can do in ten years.”
And if you think about it, the iPhone has only been around a decade, and just look how many businesses models have been disrupted and how many it has enabled. Certainly, without the ubiquity of mobility, there would be no Uber, no Airbnb.
What is clear though, is that there is no shortage of compelling use cases for voice-enabled AI in travel, and research from e-Marketer seems to indicate that, though still some way from mass adoption, people are becoming more comfortable with the technology. Those born between 1981 and 2000, the so-called millennials, are proving to be the heaviest users, rising from 23.6 million in 2016 to a projected 39.3 million by 2019. Meanwhile, Gen X’ers are expected to grow modestly from 15.6 million to 17.2 million over the next two years, while the number of Baby Boomers hover around the 10-million mark.
At some point then, can we expect a platform jump? Greif thinks so: “Historically, we have moved from the PC to the Internet and now mobile-enabled Internet. Today we can text 33 words a minute, we can type about 40 and we can speak about 140. Of course, not every use case will make sense for voice – in a meeting you might want to text quietly – but instead of putting your head down and having the friction of going in an out of apps, why not just speak in a request,” he says.
This has a potentially fundamental impact on the app world, which he argues could just disappear, not to mention traditional media like TV. It also might explain why Apple, for one, is looking, as Greif puts it “to get religion on augmented reality”. Because if Apple doesn’t have a product in the next platform jump, and that proves to be voice, or even hand, activated AR glasses then what happens to the iPhone in its current incarnation as a handheld – or, for that matter, Siri or iTunes?
The answer to these, and many others, remain swirling in the crystal ball, but whether you are a hotel, airline, cruise, rental car company, rail operator, there will be business and revenue implications. While predicting the right time to market is certainly an art, as we heard from Eurostar Head of Digital Neil Roberts yesterday, travel brands should at least be looking, as Eurostar is, at the leading digital voice assistants, and experimenting with integrating the technology. Firms should also be looking to the future to understand how younger generations in particular are using voice search. “Teenagers today are growing up with the expectation that they will be able to use voice,” stresses Greif.
So going forward, the writing is on the wall: travel brands should therefore be working very hard to give their future customers something to shout about.
At EyeforTravel North America in Las Vegas next month (19-20 Oct), Microsoft’s Stuart Greif will be speaking on a provocative panel, along with executives from Expedia and HotelTonight, about AI’s potential to revolutionise travel. More too next week from Microsoft on how augmented visualisation of CRM around the customer is the next big thing