As a concept, artificial intelligence has technically existed since the 1950s.
Specifically, the term was first coined in a conference at Dartmouth College in 1956, and has since come to be known by the more simplified initialism of AI.
It may have far future implications, but artificial intelligence is used now in more aspects of our lives than we are likely aware of – from the everyday fraud detection and shopping promotions, to more controversial systems such as facial recognition.
While we’re still longing for Marty McFly’s self tying shoes and hover-board to become part of the norm, AI is one aspect of the old sci-fi world that really has come true.
The arrival of machine learning
Back when we were dreaming of driverless cars and superhuman cyborg law enforcers, we couldn’t really comprehend how the 21st Century would shape out.
In the 80’s there was no internet, we couldn’t text each other, and GPS tracking hadn’t stretched any further than military use.
We had no idea how the development of technology would influence the world, and one of the greatest impacts on the evolution of AI has been the explosion of cloud-based machine learning services.
Subscribe to the PhocusWire newsletter below
Machine learning, in layman’s terms, is software that self-corrects and evolves its algorithms to find the optimal way to reach its goal.
As a result of this constant learning, these systems don’t need to have the “optimal way” of doing something built in from the start. The technology takes a framework of what to try, with metrics of what “good” should look like, and finds better outcomes using a structured basis of trial and error – continually improving with every attempt.
The technology sounds incredibly complex, but what makes it even more interesting is the plethora of machine learning services that are not only accessible, but also easy to use.
Perhaps disappointingly, this isn’t the once imagined sci-fi world of a few mad boffins scrambling around a giant computer with hundreds of flashing lights.
Instead, some of these services are generic, blank frameworks, where you build your own models for bespoke outcomes.
We’ve recently seen a massive growth in the use of Application Programming Interface (API) based, single-focus services, for solving specific problems, such as translating, hand-writing, identifying content in images, and even recognising faces.
Not only does machine learning exist to solve the tasks that both humans and normal applications don’t know how to solve, but it will go one step further and find the best solution that reaps the best possible result for these tasks.
However, the technology can’t be shoehorned into every situation. Tasks have to be complex, with little consistent structure, so that frequent errors can be afforded.
With that said, the travel industry can certainly benefit from machine learning more than it currently does.
Destinations at the Phocuswright Conference 2019
Hear from Klook, GetYourGuide, Certares and FourSquare. Network with many others including TourRadar, Peek, Rezdy, Brand USA, JTB Tourism and Travel Portland.
An opportunity for tour operators?
Machine learning isn’t completely foreign to the travel industry. In fact, there are a few tour operators that have been using the technology for a while now, in specific areas such as pricing and yield management.
However, it’s quite clear – looking at most website experiences – that machine learning is currently underused within travel.
In addition, combined with the likes of Robotic Process Automation (RPA), it could also be used to address significant efficiency challenges in other areas, such as resolving costly duplication and incorrect product and contract loading – removing the painful cost and bad customer experience of loading the same accommodation incorrectly multiple times.
Not only is there opportunity with regard to improving how a travel company sells, but also now in the area of making sure that what companies are selling is optimal for the customer as well.
What should travel companies do?
In the world of machine learning, there is one recurring phrase: “data, data, data…”. The more you have, the more you know, and the more you know, the higher the chance of success, both for the customer and for yourself as the selling company.
Today, there are more machine learning services available than ever before, with no need to buy any expensive hardware or implement complex software; but they all require as much data as they can get. Because of this, tour operators should be asking questions of themselves, understanding what they already know, and what else they need to find out.
Knowing this and making the information safely accessible is integral right now – especially considering the speed at which technology is developing, and with it, competition.
Technically, we can now hold all of the world’s data many times over, so there is no longer any excuse for companies not to retain all of theirs – compliance and regulation permitting, of course.
Using data, and putting it into a machine learning service can only help a company. The necessary cloud-based machine learning services are accessible and ready to use.
Gaining an advantage
We don’t need to buy specialist computers with bright flashing lights, programmable only by those too clever for NASA. Machine learning services are simple to integrate, and are easily configurable – travel companies just need to take the initiative to implement the technology.
Whether it’s looking to improve customer service, streamline bookings or to upsell with added extras, no physical man hours are required to interpret information, everything is completed by AI, and once these services are introduced, tour operators can expect to see gradual improvements.
There really is no reason why the travel industry shouldn’t be using machine learning. In such a competitive environment, the integration of what is actually quite a simple system can give travel companies a significant advantage.
About the author…
Nigel Beighton is chief information officer at Atcore.
Credit: Google News