The customer contact sector has had a new toy for some time now. It's called "chatbot." More and more companies are opting for this form of automation. However, this is at odds with the pursuit of excellent customer experience. In addition, Gartner warns that a significant portion of the chatbots that are used now can go into the trash in a few years.

At the moment, the application of chatbots in contact centers is a hype. When you visit a website, you will increasingly come across an online chat window. Is there a human or an algorithm on the other side of that screen? Often, after a few questions, it quickly becomes clear whether we are dealing with a human or a bot. Is the hype surrounding chatbots justified and is this technology a valuable addition?

Similar to FAQ

First, a look at recent articles that contribute to the high expectations surrounding chatbots. Naser Bakhshi of Deloitte Consulting gives on Emerce.nl a overview of different types of chatbots and the benefits that companies can achieve with the use of chatbots. He classifies them on the basis of 'capabilities'. The simplest bot is scripted and signals certain keywords or phrases in a conversation to which predetermined answers are given. Bakhsi likens this bot variant to presenting some kind of FAQ feature.

In second place, Deloitte sees the bots that can recognize customer intent because they are able to analyze relationships between words through machine learning. And third, there are virtual agents, who understand what a human is trying to achieve with the interaction and can hold a conversation themselves. Whether that form exists is left open, as is the definition of 'having a conversation'.

Improving processes

According to Deloitte, companies can use bots to improve their processes. The quality of customer interaction can be improved because bots are active 24/7 and always provide consistent answers. Deloitte also states that companies can achieve efficiency gains: bots take work off the hands of agents. Customers are addressed sooner and employees can focus on more complex tasks. And the third improvement opportunity Deloitte sees in the field of customer centricity; automated dialogues give companies a better view of the questions asked, the quality of the answers and what is important to consumers.

The fact that bots are active 24×7 is an advantage. Providing consistent answers seems like an advantage, until the processes or context change and a different answer is needed. With a simple and well-arranged chatbot function, making adjustments is easy to implement. The question is whether this is also the case for complex organizations that have large in-house knowledge systems.

Deloitte's claim that a chatbot contributes to efficiency and customer centricity also raises questions. If an organization already has an FAQ section on the service pages (searchable or not), a chatbot is an addition. Just like an FAQ section, a chatbot will not be able to provide exhaustive information, but the difference is that the customer is offered an extra channel where it is not visible in advance what the chatbot is and is not capable of.

This invisible limitation even applies to a chatbot that is presented as an expert in an area, such as booking airline tickets. Chatbots will regularly have to forward 'conversations' to an employee. The chance that a customer will be disappointed and that an employee has to take over a conversation is therefore high. And there goes the efficiency gain.
Research by Teleperformance shows that younger consumers in particular have no problem doing business with chatbots. But maybe you remember the heyday of the set menus? That was especially useful for the organization; the consumer didn't like it.

Customer as guinea pig

The use of technology of which it is clear in advance that it will not be successful in a large number of cases is at odds with the focus that contact center managers try to place on exceeding expectations: offering excellent service and pleasant customer experiences. Companies that use chatbots are setting the clock back about ten years in this area, as the example provided by CustomerFirst shows.

Reaal uses the chatbot of Embot.ai. There, the solution percentage at the time of the introduction was 35 percent. According to the insurer, it should gradually increase to 70 percent. Interestingly, Reaal chose to automate the 70 most frequently asked questions. The customer does not know in advance what those questions are (this is an important difference with an FAQ overview). There is therefore a good chance that the customer who enthusiastically asks a question to the chatbot will be disappointed during the dialogue.
Deloitte also says with dry eyes that when introducing a chatbot you should 'manage customer expectations – announce that they are talking to a bot and that it is not that smart yet'. Exposing the customer to this learning process raises questions to say the least. Why is the chatbot not trained behind the scenes?

Moment of truth

Companies in particular where the moment of truth plays a role, such as in the travel world, are taking a risk with the use of chatbots, according to Don Fluckinger, journalist at TechTarget. He cites Jason Baker, senior vice president of operations at Entertainment Benefits Group (EBG), which arranges discount tickets for 61 employees of 40.000 employers. The company has several contact centers.
EBG strives for "personal and memorable experiences," according to Baker. He says: “A chatbot – I understand the reason behind it, and, depending upon the type of environment, it might make sense – but in the travel and entertainment industry, you have to have the personalized touch with all interactions.”

Chatbots have so far only been suitable for conversations that amount to 'predefined click paths'. Deviation from the predefined script leads to the end of the conversation or human intervention. Also Bart van der Mark of Cognizant, speaking at Entopic's Chatbot Conference, emphasizes that bots are good at one activity – like booking a plane ticket.
His advice: “Don't make it complex. Because there are no standalone bots that fully understand the customer's intention during a normal conversation and can fully service him.”
Adfiz, the trade association of independent financial advisors, is even very conservative in its predictions: it will take another ten years before advisory robots can have full-fledged advisory conversations with customers.

In the trashcan

Chatbot technology is probably developing faster than Adfiz estimates; it is important that companies remain 'connected' to these developments. According to Gartner Research In 2022, 70 percent of customer contacts will be handled with the use of machine learning applications, chatbots or mobile messaging. By 2022, 20 percent will be handled entirely by AI. This means that considerable progress still needs to be made in the coming years.

Gartner also expects that by 2022, telephone customer contact will account for about 10 percent of all customer service interactions – although employees will still be involved in customer contact in four in ten cases. Mind you, in contrast to the current advance of chatbots, four of the ten chatbot applications launched in 2020 will have been replaced by 2018, according to the research and consultancy firm. Contact center managers who start working with chatbot technology now can therefore already inform their CFO: a second generation of chatbots will probably be available in a few years.

Blackbox

What applies to customer contact agents also applies to chatbots: the resolution rate is the most important measure of success. Companies that use chatbots should realize that they are adding a digital employee to their front office who is only good at one task, but even there is a risk of failure: chatbots are a black box whose ultimate effects on the customer experience are not clear .

Read also BAD CHATBOTS INTEGRATION BIGGEST IRRITATION

 

Technology