Chatbots are everywhere. A recent Oracle survey reported that around 80 per cent of businesses will be using them by 2020. By this time, Gartner predicts that 85 per cent of customer interactions will be handled by a machine and Bill Meisel of TMA Associates has projected that chatbots will generate over $600 billion in revenue. But what is a chatbot?
Chatbots are conversation-mimicking computer programmes that provide customers with an instant personalised response to their questions – meeting their needs immediately and saving businesses time and resource.
But would a chatbot improve your customers’ experience? And, if you are going to make an investment, what steps should you take to ensure implementation runs smoothly?
During any discussion about chatbots, you’re going to run into technological terms that require some simplification. Chatbots use the following technologies to imitate human conversations.
Artificial intelligence (AI) is the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
Machine learning is an application of AI that can access data to automatically learn and improve from experience without being explicitly programmed.
Natural Language Processing (NLP) is a branch of AI that helps computers understand human language as it’s spoken and written to be able to understand intent.
To use chatbots, you need to use messaging applications – common apps include Facebook Messenger, WhatsApp, SMS, WeChat or voice (Amazon Alexa, for example). The proliferation of such channels means that opportunities to reach customers are constantly multiplying.
Why your business needs chatbots
From speeding up response times to enhancing conversion rates and adding personality to your products, chatbots can positively impact every department of your business.
If we assume the Pareto principle holds true (the 80/20 rule), around 80 per cent of enquiries received by the average customer service team relate to around 20 per cent of the topics they cover. Even if chatbots are only programmed to handle that 20 per cent, you’ll immediately start to see customer engagement sped up, without the corresponding spend on time and resource.
Imagine your customer service chatbot [conservatively] answer 50 per cent of these frequently asked questions conversationally and to the satisfaction of your customers. Your customer service agents are now freed up to take on more complex questions and complaints that lie outside of the current abilities of AI (which will improve over time), ensuring that your customers always get the resolution they’re looking for.
And it isn’t just basic questions that your chatbot has the potential to answer. If your customer is experiencing an issue with your product, he or she can connect with a chatbot, which can then provide troubleshooting information as well as a recommendation of how to fix it. Chatbots can also share product recommendations with customers based on their personal preferences. It can even become a personal shopper – helping your customer find the perfect product.
Test, test and test again
Done right, chatbots are a fantastic way to interact with your customers and deliver industry-leading customer experience, but it’s reasonable to feel anxious about the lack of human control. So how do keep customers onside and maximise the success of the bot before you roll it out across the entire business?
Testing it out internally reduces the chance of anything going wrong when your customers start interacting with it.
The next phase of testing is to select a small sample of low-risk customers to roll the bot out to. You can then use analytics to see how the bot performed. You might want to incentivise the test group by giving them a special offer based on them completing a questionnaire about their experience with the bot.
Once live, a number of considerations need to be factored in to maximise the chances of success.
What does success look like for your bot? This may relate to existing KPIs or you might want to set up a before and after assessment. Consider things like engagement levels, goal completion rates or the number of times your chatbot has to transfer to a human for help. There’s no set rules, but by setting out clear objectives from the start, you can easily measure success in the future.
Language, slang and accents are relevant here, as is speed of talking for people less comfortable with bots. This is where NLP and machine learning comes in – able to not only detect slang and accents but also learn from them to understand intent, regardless of the specific words or phrases used.
It’s crucial to your bots success to gain customer trust right from the beginning so think about how personal information is being utilised by the bot. Access tokens are a good example and Facebook Messenger chatbots regularly use them.
Escalation to human handlers
Bots need to be able to identify unhappy customers, or those with complex queries, and ensure that information they collect is available to the person who takes the call on. This means that these customers are automatically connected to an informed agent and they don’t have to repeat themselves.
Continue to add responses and Natural Language capabilities to improve the experience for users progressively, as your bot becomes able to deal with additional intent types.
‘On brand’ bot
Don’t forget to give your bot your brand’s tone of voice but remember, there may be times when you’ll need to clarify that it’s a chatbot rather than a human.
The potential for innovation is increasing all the time, with developers working feverishly to bring new consumer experiences to market. When WhatsApp opens to bots, it will unlock direct access to over a billion new users.
Even if your business is just starting up, you’ll gain a competitive advantage and future-proof the business by investing in chatbot technology today.
Chris Crombie is product manager at Engage Hub.
Further reading on chatbots
- How UK consumers feel about AI and chatbots in customer service – survey finds that 51% of consumers prefer for their questions to be answered by chatbots with AI while they are shopping online or using an app. And nearly a third (32%) say they would prefer a working chatbot powered by AI to a human customer service agent.