Article

Tech Tuesday – How Chatbots Are Improving User Experience

24 January 2017 | Ryan Coughlan | About a 6 minute read
Tags: amazon, amazon alexa, bots, chatbots, customer service, facebook, google home


It used to be the stuff of science-fiction, but with the introduction of products such as Amazon Echo and Google Home; chatbots and virtual assistants are becoming a part of our everyday lives. Companies are focussing more and more on how they can use this technology to automate conversations and business processes. But when are Chatbots actually beneficial to your users?

So what is a Chatbot?

A Chatbot is an autonomous piece of software that conducts a conversation with a human, simulating human-to-human interaction.

Why use a Chatbot?

With the quality of A.I. within these systems improving in recent years, they can now provide a much more personalised and streamlined experience than ever before, with responses algorithmically tailored to suit certain queries more reliably and contextually than Google would produce. Therefore, a conversational approach may be easier and more efficient for a user to get that tailored experience to solve their problem than trawling through site pages to find their answers.

Chatbots have also proved to be useful in customer services, providing an alternative; almost human-like interaction without the delay of waiting for a human to answer the phone. For example, Apple could simply use a Bot to identify your problem and book you in for a visit to their store and direct you to a human only if necessary. This would maintain their ability to reach customer service KPI’s and reduce costs, whilst still providing a personable service. Using a bot will make the service more useable and accessible for all demographics.

Types of Chatbot

There are 3 main levels of complexity that chatbots can be categorised into, with each serving a purpose for differing needs:

The basic bot

Only offering limited user interactions specific to an end goal. Keeps track of the conversation and retains only basic information about the user such as name, email or phone number. Only asks closed questions so that it can steer the direction of the conversation. These are very popular in messaging apps such as Slack or Facebook’s Messenger. For example https://www.messenger.com/t/CNN/ or https://x.ai/

The Learning Bot

Interactions with this type of bot are triggered by the user who will have an end goal in mind. For example, they land on a hotel comparison site wanting to book a hotel room but needs assistance on finding relevant information.

User: I need to find a hotel in Manchester this Friday

Bot: Great, how about this (link) for £35 per night

User: I need one with a double bed

Bot: Okay, how about this one for £43 per night

User: How close is this to the city centre?

Bot: 2.4 miles, £10 in a taxi or 30 minutes walking

 

The conversation is open and able to keep track of the user’s specific requirements. It keeps track of your entire history of conversation and adapts to your preferences.

The Above and Beyond Bots:

Closely related to the learning bot this bot goes that extra mile to give you a personalised experience. This bot collects a history of the users interactions and reviews it on a regular basis. The user-bot interactions can be triggered not only by the user but also by the bot itself via push notifications or messaging apps. From these interactions it adapts and learns any underlying behaviour patterns so that it better help you. For example http://poncho.is/ is a weather chatbot integrated into Facebook Messenger and will not only give you weather updates but offer you other information such as present ideas or horoscopes.

You can utilise a Chatbot into just about any scenario, from Slack, to SMS, to being embedded on a website. Different mediums will suit different tasks. For example, on-website Bots may be the best option for handling customer service queries.

Google & the future

Google are really driving the way forward with AI and have made their Natural Language API publicly available. You can try it out here: https://cloud.google.com/natural-language/ to see how they break down the structure of your inputted sentence. Any objects you have referenced and whether the sentiment of what you said is positive or negative.


The industry is moving at a rapid rate and we are in for an exciting road ahead. Here’s a few resources we have pulled together if you are interested in a bit more research:

  • Botkit is a handy toolkit for developers to build bots for messaging platforms such as slack and facebook. This toolkit makes it fairly easy to get started and to experiment with creating your own bot. For example you could create a bot to handle your company expenses. If you have a great idea for a bot that could take some of the load off for you and your colleagues, why not give it a go?
  • If you’re an experienced developer and want a more sturdy framework for building a bot specifically for Facebook, the Facebook m Platform is perfect for your needs! Techcrunch have gone into detail about the importance for facebook bots here.
  • Microsoft have provided and maintained a development kit to build an extensive bot for all of your needs! You can read more about this here.
  • If you want some hi fidelity designs to show your customer just how effective your bot is going to be, botsociety is the place to go!

At AND Digital, we strongly believe in moving forward with the cutting edge technology, without over-engineering the problem. With a growing number of useful API’s in this field, we see great customer service prospects in this area for our clients to give a more personable and efficient feel to their service while driving costs down.

Read More From This Author

Share this blog post

Related Articles

Careers

We’re looking for bright, dynamic people to join our team!

Discover More Roles