Artificial intelligence in customer service: Five tips for implementation

The basic prerequisite for the use of AI to become a success for companies is customer acceptance. The digital consultancy elaboratum suisse has therefore developed a Study on the acceptance of artificial intelligence using the example of AI-supported online services in the financial services and insurance industry. The survey of 633 Swiss citizens showed: Young, high-income Swiss are particularly open to artificial intelligence. To increase acceptance, companies should make the initial use of AI as simple as possible and provide customers with strategic support. After all, their positive attitude toward AI increases significantly once they have already used it.

Experienced users are more open-minded

Interestingly, more than half of the consumers do not even know whether they have ever had anything to do with artificial intelligence. It seems to be difficult for the test persons to even recognize artificial intelligence in customer service on websites. As the study shows, only 28 percent of all subjects who have not yet had any contact with AI have a positive attitude toward AI. The figure for experienced users is quite different: 49 percent of them said they were open to artificial intelligence. But who are these users who already use AI and have a positive attitude toward it? The study shows that those responsible for AI services should focus on younger, male or high-income customer segments.

Most important acceptance criteria for AI: quality and data protection

But what factors increase the acceptance of artificial intelligence? According to the study, quality and data protection are the most important acceptance criteria for online services with artificial intelligence. The respondents also attach great importance to ease of use and user-friendliness. In terms of implementation, companies can establish good user guidance for a chatbot, for example, by means of predefined response options so that users do not have to interact exclusively via more complex free-text entries. Also important: Since only a few users currently trust an AI rather than a human, companies should counter user skepticism with explanations and the combination of services with a human contact as a fallback solution. Nevertheless, AI is not to be considered absolutely defeated in the question of trust. Even though the majority of users clearly trust a human more, it is astonishing that more than a third already trust AI and humans equally. Study author Marco Schulz, Director elaboratum suisse GmbH, emphasizes: "There is huge potential here for meaningful and customer-centric AI services that meet with broader acceptance."

Artificial intelligence in customer service: five tips

 What can Swiss companies learn from the study results? And what should they consider if they want to introduce artificial intelligence? elaboratum gives five tipshow the introduction can succeed:

1. No use without use: Companies should not offer an AI service just for the sake of AI. The customer does not use a service just because AI is above it. The focus must therefore be on the benefit of the service for the customer! This can be a faster processing of his inquiry or the immediate, appropriate answer of a service chatbot at 11 p.m. at night.
2. Man and machine - things work better together: Customers trust a human contact person much more than artificial intelligence. E-commerce companies should therefore convey to the customer that there is an employee at the end of the service who will check a conspicuous result again or to whom the chatbot will divert a muddled dialog in time. Co-study author Claudia Brinkmann, Senior Consultant at elaboratum, advises: "The desire of many test persons to have the option of contacting an additional human contact person in addition to the interaction with the AI should already be taken into account by companies during the requirements analysis and planning of AI-supported services."
3. AI does not replace the user experience (UX): For services that integrate AI, the same applies as for other online services and functions: Design and implement in a customer-centric way. In order to develop high-quality AI services, a combination of meaningful AI integration, holistic thinking in customer journeys, and good user experience is crucial. This combination is often not trivial, as the disciplines are already sufficiently complex when considered individually.
4. Quality and data protection - dream team for more acceptance: As the study presented above shows, quality and data protection are the most important acceptance criteria for online services with AI for the customer. Although this is also the case for services without AI, companies should take their customers' concerns seriously: For example, customers tend to react with surprise and uncertainty when a chatbot in the customer portal refers to existing contracts or purchases already made. Even if this information about customers is available in the customer portal, companies should carefully check whether all technical implementation options regarding personal data are really mandatory in the first step.
5. This product may contain traces of AI: Am I actually dealing with AI now? Almost half of all respondents in the survey stated that they often do not know whether they have already used a service with AI or not. Therefore, when using a chatbot, for example, companies should inform the customer at the beginning of the communication that they are using AI with this tool. If the customer knows that the person they are talking to is merely a machine, they are more likely to look past it if an answer sounds a bit rude or unwieldy. They are also more likely to forgive a bot for a misunderstood question. But beware: the information alone does not entitle the e-commerce company to try the customer's patience. If a customer repeats his question, the chatbot should recognize the repetition and then either admit its limited knowledge or hand over the conversation to a service employee.

The study shows that customers often have a misconception about artificial intelligence and its potential applications. Those who educate their customers and accompany them during initial use can significantly increase acceptance of their AI services. This is an important first step toward being able to exploit the potential of AI in the future.

Source: www.elaboratum.ch