Suppose you want an overview of recent expenses with your credit card company, or a travel schedule from your holiday organization, or you want to search for and book a flight with an airline, in all those cases a chatbot will do just fine. More and more organizations are therefore using chatbots to service customers. As long as it's a simple task. “Start small, select a simple task and optimize the chatbot in everything”, it was regularly said during the recently held Chatbot Conference of Entopic in Utrecht. An impression.

Chatbots, they have been around for 15 years, but are currently completely 'hot' again. A chatbot provides quick answers and is available 24/7. You can find them on websites and especially in Instant Messenger programs, such as WhatsApp, Facebook Messenger and Skype. If you are looking for contact with an organization, there is a good chance that you will be dealing with a chatbot or conversational copy.

trend

Many organizations are now familiar with this trend, but do not yet know if and how to start with it and for what purpose they can use it effectively and efficiently. That is why Entopic, an expert in the field of content, organized the first chatbot conference in the Netherlands.

“After the mobile app, you now see that the chatbot presents itself prominently as a means to properly serve your target group,” explains Niels Vonk, Managing Director at Entopic. “Organizations need specialist information. How do you get started, what are the success cases, what are the prospects?”

“We have the knowledge, the network, know the speakers and are now bringing that together for the first time during this event. Finance, retail, government and energy companies, many large brands are focusing on chatbots. Customers are more and more used to using chatbots. At Coolblue for example. As a result, they expect to encounter a chatbot at other organizations as well.”

Self-learning bots

The new generation of chatbots have long since passed the 'fully scripted' stage. They are self-learning and improve with every new contact based on Artificial Intelligence. A good example of this is the Transavia chatbot. It serves one purpose: to book airline tickets. You can communicate with it in all sorts of ways in natural language.

“But,” assures Cognizant's Bart van der Mark, “this bot, like all other bots, is not comparable to human intelligence. You can book a ticket with it, but nothing more. And it does that very well. My advice: take that as an example.”

“Select one activity, provide a chatbot plan and put together a roadmap for a pilot. 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.”

Human service

“For flawless communication, human service is indispensable. Good cooperation between the chatbot and the service employee is therefore very important. Because then the employee can quickly take over the conversation if it threatens to get stuck.”

“Make sure that the chatbot and the service agent know from each other who is doing what. Only then can you keep a chatbot afloat. Its added value? Lower costs for customer service, higher customer satisfaction and also employee satisfaction, because routine jobs are automated, and more sales revenue thanks to personalized digital service.”

Artificial Intelligence.

It should come as no surprise that customers are in principle open to a conversation with a chatbot. Because doesn't every search for information start online? There is no service agent involved. Research organization Gartner predicts that no less than 2021 percent of all customer communication will be via AI by 15.

Currently, the majority of all customer questions are handled by assisted service. But the self-service share will soon surpass this. Customer service strategist Martin Hill of Brainfood shares this conviction: “A well-functioning chatbot searches for specific information, does not send you complete PDFs, but selects information based on customer needs. Customers keep asking the same questions. AI can figure that out. With that knowledge, customer journeys and self-service can be improved and proactive service can be offered more often at relevant times.”

Hill also knows that self-service is not an option if the customer journey is complex. The challenge in that case is to make switching between bot and human as easy as possible for the customer. “Promote the bot, just like Oxxio does for example, and clearly indicate to the customer through which contact channels the customer can also obtain information, if the chatbot does not provide him with sufficient service.”

Natural language

What helps the adoption of chatbots is the ability to communicate in a natural way. Just ask the drive-through burger chain. Customers express themselves in the most creative way to order the beloved burger: 'The Big One', 'Biggie', that Big Fat Monster', et cetera, et cetera. And in all cases, the chatbot in question knows that it is about that double hamburger.

A good understanding of the customer's intentions is a prerequisite for a successful conversation. “Many developers write boring scripts,” says Hans van Dam of Robocopy. “They have no insight into how brains work. Psychologists do; we add them to our team of script writers.”

A successful conversation has a number of characteristics: personality, natural language, empathy, helpfulness and persuasiveness. “Focus on key customer journeys. Ask yourself: how does the customer feel now and how do you want them to feel? Let that be the starting point of your natural language. Then adjust your language to personas that you use when approaching customers, for example.”

“Don't leave the customer in the dark, make conversations simple, acknowledge their troubled feelings, show understanding and say sorry – even if the in-house lawyer is afraid of a claim. Provide connection and build empathy loops into the conversation if you know you're asking the customer for what is necessary. And finally: don't overwhelm the customer with too many options. With a chatbot of an insurance company, we reduced 7 choices to 3. The conversion increased from 12 to 64 percent.”

Simplicity

Simplicity, customers love it. Eneco also experienced this at the end of 2016. At the request of customers, the energy company built a chatbot in Messenger to transmit meter readings. A child can do the laundry. It is a process with a head and a tail.

Building, testing with a thousand customers, making improvements, until the bot went live took only 3 months. Employees were informed about the chatbot in a fun way, customers shared their enthusiasm via social channels and the media generously covered the news about the new tool.

Remarkable: the average user is between 50 and 55 years old. Eneco will continue on this path and wants to help customers 'get the most out of your smart home through digital conversations'. In addition, the customer will talk to a chatbot more often than now than type with it, probably via Google Home, which is catching up in the Dutch language area, or Alexa.

Customer journeys

Based on customer questions, insurance company asr led the development of a conversational website that revolves entirely around customer questions. The homepage of this insurer shows a large chat window. This enables the organization to find out in record time which questions customers are asking. The knowledge base is enriched with up-to-date knowledge of customer contact advisors and data from call reasons.

Depending on the request, the organization shows the chatbot or routes it to another contact channel. As a result, the loss of contacts can be kept to a minimum. Not all five thousand financial advisors the firm works with are ready for this digital way of working. But now that the insurer's opening site also has categories in addition to a search bar, there is much less cold feet. The use of conversational copy has increased by 60 percent.

A big advantage for the organization is that the chatbot forces you to think in terms of customer journeys. Nick Aanraad: “Call it the other way around. We do not publish loads of information that we hope customers read, but start from the specific customer journey and check what we can offer exactly. We proactively inform customers about what is to come. This enables us to solve relevant customer issues quickly and effectively. Now with the chatbot, later with live chat.”

Read also RPA IN HR: CAN WE DO IT WITHOUT IT?

Customer Experience, Technology