EOS 2021 Chatbot Survey: Two of three companies already use bots.

The digital helpers, which are especially popular in the e-commerce sector, but also with insurance companies, banks, and energy and telecommunications service providers, handle around a third of the digital communications between companies and their customers. And wherever chatbots are used, they already independently resolve nearly half of the issues communicated by customers. Chatbots can provide advice to online shoppers or help in the selection of an electricity or cell phone contract, supply contact data and account numbers, give information about the delivery status of pizza and packages, and collect customer data. Yet despite the wide range of possible uses for them along the customer journey, many of them are still mainly used for the initial contact (97 percent), followed by customer service (51 percent) and product advice (39 percent). 

Russia is the exception: The fewest chatbots are used there, but their use is much more variable than in all other countries in the survey. “Quality over quantity” could describe the situation here. More than 70 percent of the chatbots are already being used there for provision of product advice (78 percent), customer service (73 percent), and collection and analysis of customer data (71 percent). Especially when it comes to generating data, the Russian bots are way ahead of the rest: “Collection and analysis of data” is seen by 64 percent of survey respondents as a big advantage of a chatbot – but only 29 percent of the companies have done this in practice using chatbots so far.

The survey participants also see the untapped potential: Of the companies already using chatbots, 88 percent would like to develop them further in the future. Of these, 26 percent are planning a qualitative optimization in the form of an increase in the intelligence level. About half as many companies (15 percent) are focussing more intently on quantity, i.e., increasing scalability or the volume of inquiries. However, at 47 percent, the greatest share of the companies sees optimization potential in terms of both quality and quantity.

Utilization of the different channels on which the chatbot technology can be used can also be improved. At 74 percent, most of the digital communication aids can currently be found on company homepages (on-page chats). This is followed at some distance by in-app chats (24 percent), Facebook Messenger (18 percent), and WhatsApp (17 percent).

Most of the surveyed companies are of the opinion that the importance of chatbots in customer communications has increased during the corona pandemic. Asked about user acceptance, companies already using chatbots give a clear answer: 96 percent of them say that users accept chatbots. Despite this unequivocal statement from chatbot-using companies, the technology is irrelevant for a fifth of participants, who currently have no chatbots and are not planning on introducing any.

Participating companies view the two greatest hurdles by far to be the extensive data protection requirements (70%) and the revision of in-house database structures necessary for introduction of chatbots (59%). In contrast, the investment needed for the technology hardly plays a role: Only three percent of the companies that are not planning any chatbots state excessive costs as the reason for not using the technology. 

Belonging to an industry deemed too sensitive, in contrast, does not present a problem in the use of a chatbot, as EOS itself shows: The digital assistants are already being used successfully by the technology-based debt collection service provider in Belgium, Croatia, and France. “Since the rollout of our chatbot Tom, calls to the service center have declined,” says Wesley van de Walle, Project Manager at the Belgian EOS national subsidariary EOS Contentia. On average, the chatbot conducts more than 300 conversations with defaulting payers every month. That’s no wonder, since every tenth survey participant sees anonymous communication as the most important advantage for customers. At first, Tom was only capable of providing basic information such as account numbers or telephone numbers and email addresses of contacts, “which, by the way, should not be underestimated and has already brought great relief,” says Wesley. Since he and his team gave Tom an update, Tom has been able to do much more: Connected to the debt collection system from EOS Contentia, the chatbot can now also process payment defaulters’ personal data. 

“After users have authenticated themselves, Tom can now, on request, give information about the amount of debt, the due date of the next payment, or any costs incurred,” says Wesley. Lea, Tom’s chatbot colleague in Croatia, also has these extended functionalities. “Meeting the strict data protection requirements is a matter of course for EOS – the challenge in developing chatbots in our business with sensitive data is to find a balance between ease of operation and the necessary security,” says Franjo Glibo, Lea Manager and Head of the IT Department at EOS in Croatia. Franjo sees advantages for users if they divulge personal data to a chatbot. “Companies should communicate openly and point out what consumers get out of sharing data with them.” In this way, both sides could profit. “EOS can, for example, use data to achieve the best possible individualized receivables management – that even has advantages for defaulting payers,” says Franjo. He and his team are currently working on making it possible for defaulting payers to upload payment documentation such as bank statements to the debt collection system with the help of Lea. Another thing that should be introduced is a direct transition to human service staff when Lea comes up against her limits. 

“This happens, for example, when a defaulting payer wants to negotiate an individualized payment plan,” says Franjo. This and similar tasks remain too complex for a chatbot for the time being. EOS employees therefore do not need to be afraid that Lea is going to replace them – instead, they can apply themselves to the truly challenging cases, and that’s a lot more satisfying than reciting the account number for the tenth time, says Franjo. Michaela Homann, Head of Customer Communications at EOS Technology Solutions (TS) in Germany, is following Lea’s development very closely: “In the future, we are planning to implement chatbots on various channels in Germany as well. We can definitely benefit from the technological basis and the knowledge gained in Croatia and Belgium,” says Michaela. For this, e.g. the Croatian requirements document could be adapted and translated into German. Besides EOS in Germany, the EOS national subsidiaries in Spain and some of the Eastern European countries are also reviewing the possibility of or planning the introduction of chatbots.

Wesley and Franjo are already also occupying themselves with voice bots. “It will still be a few years before they are fully implemented – in order to be ready at that time, we have to deal with the technology now,” says Franjo. He assumes that chat and later also voice bots will become standard tools in customer communications. The companies surveyed in the EOS 2021 Chatbot Survey think so too, with 61 percent of them firmly believing that in the long term, every large company will have to offer a chatbot. The analyst firm Research and Market supports this view: The chatbot market is expected to grow from USD 2.6 billion in 2019 to USD 9.4 billion by 2024, the experts write in a market forecast. The age of the bots has begun.