Better and automated customer dialogue with chatbots

By Fredrik Broch Elgaaen, strategic consultant, technologist and business developer, Knowit Experience

Our clients are already using artificial intelligence when communicating with their customers in a chat interface. There is increasing focus on the digitalisation and automation of processes. This is why chatbots are on everyone’s lips these days – precisely because they are a way of digitalising and automating the customer dialogue.

The yardstick is a global one, and customers expect to be acknowledged within seconds. Flawless solutions make for the best customer experience. That is why you have to place the bar at the right height and not jump where everyone else jumps.

Personalisation will soon become hyper-personalisation
Investment in big data and AI must be part of a broader picture. Knowit advises its clients to bear this in mind when designing infrastructure and other systems. If you let this way of thinking permeate the entire organisation, the path to exploiting the value that the data holds is very short indeed – not only to today’s business problems but those of the future, too.

In terms of the customer journey, we believe that the future for many sectors will be personal assistants. The way in which we interact with websites and apps will be robustly challenged by message-based interaction.

Big data, artificial intelligence and chatbots will all be at the heart of this trend. Big data as a basis for all calculations and insights, artificial intelligence that makes decisions for the company, that advises customers and personalises the interaction with each customer. The chatbot is the window to the customer.

Customer dialogue has already taken big strides towards chat and messaging services, while speech is on its way out. In the wake of this trend, there are great opportunities opening up for adopting new and exciting technologies to automate parts of the customer dialogue.

We have now reached a point where messaging platforms have overtaken the use of social media. In other words, customers are spending more time on Messenger, WhatsApp, Skype, etc. than on Facebook and Twitter. It is also important to remember that digitalisation is as important to customers as it is to companies. When customers go digital, businesses have to follow. It is all about being where your customers are.

What does it take to succeed with a chatbot?
The worst thing you can do is believe that advanced technology is the path to success. The smartest bot is not necessarily the best. The objective is to solve a task more easily, quicker and better than an alternative process. It is all about the customer experience.

How do you create a good customer experience? With customer insight. You have to know where the customer wants to find the bot, which questions they want answers to, how they feel about being assisted by a machine rather than a human, and how they want the bot to communicate with them.

”It is important to look after your customers in the right way.
Humbly admit that the bot did not understand and encourage the customer to rephrase their question or be seamlessly transferred to a human customer service advisor.”

You must instil the right expectations in the customer
Be open and say that this is a machine. Creating a bot that can answer absolutely everything is a tall order. Nor should it be your goal in the immediate future. But what do you do when the bot is unable to provide an answer? It is important to look after your customers in the right way. Humbly admit that the bot did not understand and encourage the customer to rephrase their question or be seamlessly transferred to a human customer service advisor.

Deciding what your needs are is crucial on the path to success. What do you want the bot to do for you? It is important not to bite off more than you can chew. The first tasks you pick should be simple, standardised and occur regularly. They must be tasks that a bot can solve for the customer in a way that is satisfactory for both parties. One good example is FAQs.

Finally, there are certain technical decisions to make. Which internal systems will your chatbot rely on to solve the tasks it has been assigned? There is a multitude of tools available for building chatbots. Important things to bear in mind when choosing a tool are language support and integration with different platforms.

Chatbots do not have to be a major investment
The most important groundwork when developing a chatbot is done before you even start. If you have the answers to the questions posed in the previous paragraph, you are well on your way. One good thing about chatbots is that they are easy to launch without having to make a huge investment. Once the bot framework is in place, you can build the bot’s knowledge bit by bit.

As with all automation, one argument for using a chatbot is, of course, streamlining and cost savings. Do not forget that a chatbot will be the public face of your company. In order for a chatbot investment to pay for itself, your focus must be on the customer. Only when your customers start using the chatbot service rather than your traditional customer service channels will you begin to see returns on your investment.

What does the future hold?
Gartner suggests that 90% of customer queries will take place by text by 2020. We are not quite there yet, but it is clear that text is well on its way to replacing speech. Servion Global Solutions, meanwhile, suggests that artificial intelligence (including chatbots) will account for 95% of all customer dialogue by 2025.

The way we use apps is changing. In the future it will not be about the experiences you manage to create inside an app, but outside it. In a sea of apps, a pattern has begun to emerge in terms of how we interact with them. Rather than physically having to open each app, we are increasingly beginning to interact with them from a “central command” such as a messaging platform.

The way we navigate websites will also be facing tough competition. Where we currently physically visit a website before browsing it, the future will see us requesting details of all men’s shoes in size 44 being sent to our inbox.

Nobody can predict the future, but one thing we do know: if you want to stay relevant, you have to be where your customers are.

What exactly is a chatbot?
At the bottom end of the scale are command-based bots. You write a command that triggers a rule in the back end system, which in turns gives you a predefined answer. At the other end you have self-learning bots that use artificial intelligence to interpret what is being asked before giving a suitable answer.

In terms of technology, it would be sensible to aim for something in between the two at this stage. Command-based bots are extremely limited, while self-learning bots have shown themselves to be moving in an unwanted direction. If you combine the ability to learn and understand what the customer is saying with predefined answers in a decision tree or conversation tree, you get a bot that is more than capable of dealing with most customer queries.

Knowit has the methodology to efficiently help its clients create a better and more personalised customer dialogue. Our clients can now use artificial intelligence, rules and business logic when communicating with their customers in a chat interface.