The Data Analytics Blog

Our news and views relating to Data Analytics, Big Data, Machine Learning, and the world of Credit.

All Posts

What Are Bots?

December 8, 2016 at 9:43 AM

In this blog post I’ll be covering what bots are, how bots are used, the growing popularity of bots and the three types of bots that I have come across.

How we got to bots

It’s hard not to notice the developments that have been sweeping across the analytics space over the last couple of years – big data and machine learning are two classic examples.  

The phenomenal growth in storage and processing powers brought on by big data and machine learning has made it possible for computers to not only think, but to even out-think and out-perform humans. A great real world example is IBM’s Watson, which in 2011 out-performed the two top contenders on Jeopardy.  

Watson was a combination of ground-breaking predictive analytics, combined with big data – i.e. crunching and learning from large data sources, such as encyclopaedias, dictionaries, thesauri, newswire articles and literary works.  Watson could process 500 gigabytes - the equivalent of a million books - per second.  Watch Watson taking on Ken Jennings and Brad Rutter. It is truly impressive (22m50s is where the action takes place. Although, the whole clip is worth watching). 

Computers performing like humans is not a new thing by any means.  We have for a very long time toyed with the concept of artificial systems behaving like humans - in reality and in fiction.  Who wouldn’t like their own personal assistant that knows what you like, saves you time and feeds you the right information throughout the day?  Enter the bot.  

What is a bot?

There are various definitions of what a bot is, but in our view it is simply an automated system that does something, i.e. it has some pre-defined purpose.  

Still confused by what a bot is? Then let’s try clarifying with a question. Is something an automated system that does what a human did previously?  If the answer to this question is a ‘yes’ then we are probably talking about a bot. If it moves, then it is not a bot.  

An automated dialler that selects the next best number to dial? That’s a bot.  An automated financial advisor to help you select the best portfolio based on your profile? That’s a bot.  How about a service that processes your request from a chat portal and books a restaurant for you? Yep, it’s a bot. 

There are Three Types of Bots

To categorise bots further, there are three types: 1) interactive, 2) pure process automation, and 3) coaching bots.  

1) The Process Automation Bot

Process automation bots can replace, where possible, mundane or repetitive tasks that humans do, like transferring information from a spreadsheet into a different system.  They can take a bit of time to set up and mimic the required process, but then they are good to go and can operate 24/7 with a near zero error rate, assuming the inputs remain consistent.

2) The Interactive – or “chat” - Bot

Interactive bots (“chatbots”, also called conversational UI’s) are the ones that are taking the world by storm.  Being able to hold a decent two-way conversation is one of the key features and big challenges of a good AI system.  Apart from the need to accurately convert voice to text and dealing with all the nuances that come with the spoken language, the system needs to extract the essence of what is being said, taking into account all the different ways that humans say the same thing.  And then there is the compilation of an appropriate response which can also be complex.  The execution is the relatively easy bit, e.g. the act of booking a restaurant or paying a bill.  The chatbot community has made great strides in getting the interaction bit right and the experience is getting closer to talking to a real human – some of the better bots are uncanny which can either leave you feeling suitably impressed or a bit uncomfortable.  Try this one for a demo.  See if you break it, I struggled.

There is another reason why chatbots are receiving so much focus.  For the first time ever, people are spending more time in their instant messaging platforms than on social media sites.  People don’t want to download an app anymore; they often either take too long to set up and figure out or are too clunky to interact with.  Everyone is familiar with interacting on Whatsapp or Facebook Messenger, so it would make a lot of sense if one can use this platform to get everyday things done. 

Look at what WeChat has achieved in this area.  Most users in China where WeChat is the platform of choice spend almost all of their time in this platform, chatting, reserving flights, restaurants, tracking their daily fitness and paying their water bill to name but a few.  Businesses now have the ability to set up a system that can interact with individuals (and customers), get them what they need and get a gold star in the process, all at the fraction of the cost of setting up and running a call centre.  Facebook Messenger has seen the writing on the wall and has invested in their chatbot platform, probably just in the nick of time.  

The other important point to note is that interacting with a machine is becoming more acceptable.  It is rapidly becoming quite normal to interact with a bot, especially if it will give you the right (data driven) answer in a fraction of the time, any time of day and in a familiar way.  There are still many improvements to be made in this area, but the pieces of the puzzle are starting to click into place. 

3) The Coaching Bot

And finally, keeping what we think is the best for last – the coaching bot.  Sometimes you just can’t beat that human-to-human connection BUT a machine can help you make a better decision, based on past data of course.  In many areas, face to face conversations with a real-life human agent is still the expectation, but there is definitely a role to play for data driven insights to support humans.  Applied and trained effectively, a machine learning system can make better decisions than a human can across complex choices and evaluations like 1) the best tone to adopt for each interaction, 2) determine interest in additional products and which products those are likely to be and 3) the sensitivity to pricing at an individual level. 

Is your business ready for bots?

The world of bots is still at the early stages of development and take-up, but it is growing as rapidly as big data and machine learning.  As a business, it would be unwise to ignore how bots can be used to benefit your marketing, customer relationship and operations endeavours.  One thing is for sure, how we as customers interact with our service providers will be very different in 5 years’ time. 

We at Principa are very excited to be able to support the data driven part of bots, especially to support call centre agents in making better decisions.

Get in touch with us if you'd like to discuss bots! Even if it's just a philosophical one over a cup of coffee! Learn more about our Coaching Bot for Call Centre Agents

Robin Davies
Robin Davies
Robin Davies is the Head of Product Development at Principa. Robin’s team packages complex concepts into easy-to-use products that help our clients to lift their business in often unexpected ways.

Latest Posts

13 South African Data Science & Machine Learning Meetup Groups To Join

Meetup (www.meetup.com) was founded in 2002 and is an online service used to organise groups that host in-person events for people with similar interests. A social networking tool, unlike any other, Meetup, has more than 35 million users in 180 countries. In South Africa, there are many meetups organised for various interests. Of course, our interests span data science, machine learning and AI, so we took the time to put together a (non-comprehensive) list of the groups who organise meetups that we (and hopefully you) find of interest. We hope you make it out to one of these groups meetups soon!