It seems ages ago now, but those four days off at Easter meant I could lay on a proper Easter egg hunt for my grandson in the garden. My determination to make it a proper adventure often ends up with eggs in places he’s too short to find, which has its benefits when I come to find them weeks later.

Playing the Easter bunny reminds me that I have an imaginary friend myself named after another famous fictional character. Meet, ahem, Roger Robot.

Roger Robot is my go-to character to explain how machine learning can drive better experiences. I believe it’s the branch of artificial intelligence which will drive the biggest improvements in our industry.

In case you’re confused about what machine learning is and how it differs from AI, let’s elaborate. AI means computer systems that can perform tasks in the same way as humans: like languages or making decisions based on analysing data and patterns.

Machine learning is a sub-set of AI. It trains a computer system to self-learn and improve from the data it receives, pushing the system to deliver a better service. 

This is where my friend Roger Robot comes in. He’s a hypothetical machine learning-based travel assistant, who can progressively offer better-targeted online booking options to your travellers. 

Roger is a bit clueless at first, but you solve that by teaching him to recognise different types of business travel arrangements. Let’s say your travel policy only allows economy flights but is more generous about where employees can stay. You start by showing Roger lots of economy flights and luxury hotels. Each time you show him an economy flight, you say, “This is an economy flight,” and each time you show him a luxury hotel, you say, “This is a luxury hotel.”

Being the bright bunny he is, Roger starts to notice patterns and refine his suggestions. He sees that economy flights are usually cheaper but have less legroom, while luxury hotels offer premium amenities and spacious rooms. Over time, Roger gets better at telling the difference between economy flights and luxury hotels. This process of learning from examples is what we call “training”.

Like humans, Roger isn’t perfect. Sometimes he might mistake premium economy for standard economy or a boutique hotel for a luxury one. To improve his accuracy, you keep showing him more examples and correcting his mistakes. This continuous process is what makes machine learning so powerful. 

You eventually end up with a booking tool that offers very smart, compliant choices to travellers. In which case everyone’s a winner, and it’s chocolate eggs all round. Well, enough “rabbiting on” from me. That’s all, folks: Roger and out.

Scott Wylie is Chief Technology Officer of TripStax