THE BLOG

What Is A Bot?

12/10/2016 11:59

Ask anyone in the tech industry what the next big thing is and they are very likely to say bots, artificial intelligence or machine learning. Recent advancements in innovation across multiple industries and disciplines have created mass interest, and major investment, in bots - but what exactly are they and what is their full potential?

A bot, in essence, is intelligence that is embedded on some form of hardware, be it a Raspberry Pi, a smartphone, in the cloud or in an electronic car. For example, when a smartphone suggests apps to its user based on their location or if they've just plugged their headphones in, that's the activity of a bot. We've all seen the online marketplace adverts following us around the internet offering us the same item(s) we've just bought - that's also a bot, just not a very clever one.

In order for bots to perform the desired actions, they must be trained using data - the more data, the better the bot. This is where data scientists come in; we develop an understanding of the data by exploration, determining its strong and weak points. Mathematical algorithms are then used to create models that represent the data and allow bots to make their predictions.

Sophisticated bots are intended to make our every-day tech easier to use and as tailored to our individual needs as possible. If we take video-on-demand services for example, every piece of content offered is assigned a list of characteristics that signpost the type of show it is. A bot uses these characteristics to help identify trends in content preference, which are then used to suggest relevant shows for individual users. The bot also goes a step further - it helps to determine which new shows are bought, or even created, by the services in order to satisfy, and grow, their user base.

Chatbots

The latest trend in the world of bots is the chatbot. Some truly ground-breaking work has made "useable" chatbots possible. The advancement of deep learning and computing technology has created a step-change in natural language processing, the backbone of all chatbots.

Ask a question in text or speech and, if the chatbot has been trained to handle that type of question, out comes a human-like response. They can handle simple conversations as well. However, the interaction that chatbots currently offer is still very superficial as we don't yet have the capability to truly understand the full benefit this technology could bring. Chatbots will have reached their full potential when they can pass The Turing Test - a person messaging a computer (and the bot within it) whilst convinced they are speaking to another person - but we're a long way off that.

Travel bots

If we look at the travel industry, bots are already ubiquitous - whether it's hotel bookings, train travel or taxi hailing (to name just a few). A good example of a bot in action, is that used by city bike hire schemes. The schemes use a bot to learn from the travel patterns of the bike users so it knows where to send drivers to pick up extra bike, before docking bays fill up, and distribute them to areas where a high volume of demand is anticipated later that day.

At Trainline, our bot story has flourished with the release of BusyBot - a bot that tells our app users how busy their train is likely to be and which part of the train should have the most seats available before it pulls into their platform. To help 'teach BusyBot', we've been crowdsourcing data - asking our users to share feedback on how busy their trains are - and the response has been incredible: to date over 1 million pieces of feedback have been shared with us, resulting in a very smart bot with the potential to make journeys much more seamless.

The future of bots

The biggest bot on the (not so distant) horizon is the one powering self-driving cars. This advancement is a level of modal innovation not seen since the mass-production of cars in the 20th Century and is a game-changer for multiple industries, from travel to food delivery and beyond. This innovation will change how we live our lives and how our cities operate.

We're only just scratching the surface of what can be achieved with bot technology and are set to see some very exciting developments in the coming years and decades. At the very least, it's clear from the early iterations in use today that bots have an important role to play in making technology even more intuitive and invaluable to us all.

Comments

CONVERSATIONS