Chatbots. The Future or 'All Talk'?

Although chatbots have been around in some form or another for years, the hype cycle around them truly began when Facebook announced their bot platform for Facebook Messenger just over a year ago. Facebook recently demonstrated their commitment to chatbots at this year’s F8 developer conference. Since the release of Facebook's bot platform, thousands of companies have had a go at releasing a chatbot, experimenting with how it best connects them to their customers.

There’s been a lot of developer interest in exploring the capabilities of chatbots (also known as conversational interfaces) - backed up by the rise of development platforms and tools to help build them. The usual suspects (read Google, Amazon, Microsoft and IBM) are also investing heavily in this area and opening up some of their artificial intelligence muscle to developers in order to help fuel growth.

To be clear, chatbots aren’t just available for Facebook Messenger. However it’s Messenger and other popular messaging platforms such as WhatsApp, Slack, Telegram, Kik and Skype where most of the current interest is taking place.

To date though, the actual experience of using a chatbot hasn’t lived up to the hype. So why is there so much interest in this technology and how can it best be used?

What’s a Chatbot?

Before chatbots there were bots. A bot is a piece of software designed to automate a specific task. There’s generally little to none human interaction with a bot outside of the programming commands a software developer has tasked it with.

A chatbot does exactly the same thing as a bot with the difference being it allows humans to interact with it in a conversational way (by asking it to do things via text or voice). In some cases it may use a little artificial intelligence to help interpret the meaning of what you say and to keep the conversation flowing. For example, you might ask a weather bot ‘How’s the weather looking in Melbourne tomorrow?’ followed by ‘And how about Friday?  Is it going to rain?’.

Your first interaction with a chatbot might have been the often misunderstood Microsoft 'Clippy'.

Your first interaction with a chatbot might have been the often misunderstood Microsoft 'Clippy'.

To keep things simple we can break down modern day chatbots into two major strands;

  1. Single purpose messaging chatbots released by companies on various messaging platforms such as Facebook Messenger, Telegram, Slack and Skype.

  2. Virtual assistants such as Siri, Cortana and the Amazon Alexa product range which try to assist you with any general question you throw at it.

This article aims to look at the former - single purpose messaging chatbots built for release on various messaging platforms.

Why Chatbots?

There are some compelling reasons justifying the interest in chatbots;

  • If we are to believe Gartner, by 2020 the average person will have more conversations with bots than with their spouse!

  • Messaging platforms such as Facebook Messenger and WhatsApp each have 1.2 billion monthly active users. Outside of Facebook proper they are the western world's biggest social networks. This represents an opportunity for brands and businesses to reach a potentially huge and engaged audience.

  • Chatbots are relatively easy to use. The interface is familiar - hardly any different to messaging a friend, so there isn’t the learning curve which comes with using individual apps and websites.

  • App fatigue. With people downloading less mobile apps than ever before, chatbots are a new way to reach your audience which doesn’t require the user to download anything.

  • Development costs. Once your chatbot is built for Messenger it is available across web, tablet and mobile on all operating systems (such as PC, Mac, Android or iOS). Anyone who has built an app will understand the pain and cost of maintaining it across multiple platforms. Chatbot development is relatively cheap.

  • Conversational interfaces are becoming increasingly familiar and popular - especially with millennials who are entering the stage of their lives where they will be the biggest spending demographic. In general, they prefer messaging to email and expect to receive answers close to instantly - no matter what time of day.

Ecommerce chatbot - chatShopper.

Ecommerce chatbot - chatShopper.

WeChat - a popular messaging platform in China with 846 million monthly active users could be an indicator of where things are heading. Third party chatbots have been an integral part of the platform for a few years now and proven to be extremely popular. Many tasks which Western countries still traditionally use websites and apps for are being carried out via messaging. Customer service, transferring money, ordering food, shopping and playing games is  often conducted through chatbot interaction in WeChat.

Although nothing to date has been announced, it’s hard to imagine that Apple won’t open up their iMessage platform to third party chatbots at some point. This will create another potential massive audience and fuel further development growth.

Use Cases for Business to Consumer Bots

Automation is the type of task chatbots excel at. Businesses, whether they are B2C or B2B should explore if there are high volume or repetitive tasks which a bot could automate to improve customer experience, efficiency or reduce operational costs.

Chatbots of today work best when they specialise at a particular task. Popular use cases include but are not limited to;

  • Customer service - Having 24/7 available customer representatives who can deal with thousands of customer requests simultaneously would be extremely expensive and not practical for most businesses. A single chatbot can solve this issue for businesses large and small - also having the ability to pass through messages to a real person if it can’t answer the question itself (providing the functionality of a call centre at a fraction of the cost).

  • E-commerce - A chatbot can offer personalised and guided ways to help shoppers find items they want by asking the right questions, making suggestions and taking users on a journey which would be difficult to navigate on a website or app.

  • Marketing - Chatbots represent another digital channel with unique and creative marketing campaign potential or avenue to increase engagement. Given the user data gathered on platforms such as Messenger, valuable demographic info can be used to enhance the conversational experience. A conversation or promotional offer can be hyper-tailored and localised by the chatbot depending on who they are speaking with.

  • News - Personalised news can be subscribed to and sent to users via message each day. Breaking stories and sport score alerts can also be messaged.

  • Data gathering - Conversing with a chatbot could be a more compelling way to complete surveys or gather information in bite size chunks by planting questions into the conversational flow while completing other tasks for the user.

  • Health - Chatbots can be an engaging way for people to ask medical or nutritional related questions, or even to be used as companion to check in on your mental health.

US immigration attorney chatbot - Visabot.

US immigration attorney chatbot - Visabot.

Chatbots In The Workplace

Third party developed bots are at the heart of work team messaging platform Slack. Chatbots on the platform can be used by employees to give reminders, check their calendar, integrate and pull information from other software such as Salesforce or Google Drive, book an Uber or remind you to drink more water.

Moving outside of messaging platforms - expect the enterprise to be huge area for the growth of chatbots. The ability to use natural language via voice or typing and have a centralised single ‘Siri for your business’ assistant would probably be the holy grail. Imagine having a report generated for you merely by asking ‘sales for last quarter broken down by region’ or ‘create John Doe as a new lead in Salesforce’.

Chatbots in the workplace could also assist with research, file management, staff training and scheduling meetings. Every employee having access to a virtual assistant and 'second brain' augmenting their workplace knowledge would be of value to a business.

Bot to bot communication is also an area picking up steam. If the chatbot you’re talking to can’t answer a question, it will ask another bot specialising in that area to see if it can help, sending the answer back to the bot you’re communicating with. (For the techies, this can be seen as a type of modern day asynchronous API).

Room for Improvement

If you’ve used a chatbot over the past year you may have been underwhelmed with the experience. Quite often the bots only understand very simple commands, only responding as expected if very specific language is used and leaving you hanging when they don’t understand.

Other chatbots try and tackle too much rather than trying to execute one task well. And some services just shouldn’t be bots - a conversational interface might not be the quickest or best way for a user to get the info they need.

(Chatbot fails...)

(Chatbot fails...)

There are many different linguistic ways a person can ask the same question or give an answer. Many chatbots struggle with this. This situation is improving though.  The best chatbots are starting to take advantage of ‘natural language processing’ (NLP) tools which use machine learning and artificial intelligence to better understand and interpret the many ways humans have a conversation. Microsoft, IBM, Amazon and Google amongst others have released developer APIs to assist chatbot developers with this.

One of the biggest issues with chatbots at the moment though is discoverability. Outside of the normal ways a business can promote a digital product such as advertising and providing links, there is currently no easy way to discover chatbots. There is no equivalent of an app store. Fortunately this will soon change as Facebook have announced a ‘Discover’ tab within Messenger to help solve that very issue.

The future or ‘all talk’?

Although chatbots are in their infancy with regards to potential capabilities, they already do a great job in certain use cases and their conversational skills are quickly improving. They are relatively cheap to prototype, develop and iterate on compared to other technologies such as apps and offer a familiar yet unique customer experience. Chatbots aren't always the answer but they are definitely here to stay.

This article has just scratched the surface on how chatbots and other conversational interfaces will become part of the everyday experience. It’s time to start exploring how your business can utilise them to increase customer engagement, improve customer experience and add value.