NEW RESEARCH | Customers love innovative AI services – but are afraid to use them

Consumers are more satisfied than ever with advanced AI services. At the same time, they are less and less likely to actually use them. This is according to a doctoral thesis at Linköping University.

– The more innovative an AI service is perceived to be, the more satisfied customers are – but the less likely they are to adopt it. This is a paradoxical result, says Joanna Pilawa, who has studied how people perceive and react to service innovations at a time when artificial intelligence is rapidly changing the service landscape.

‘Signals’ that influence perceptions of the company

In her study, she uses signaling theory to describe how service innovations are perceived by the recipient. When a service changes – whether it is a parcel box, a chatbot or a digital assistant – the company sends a “signal” that the customer interprets. That interpretation affects the perception of both the service and the company. It can go either way.

– When many companies introduced early chatbots to streamline customer service, customers didn’t see it as an improvement, but as an annoyance.

So companies can control the innovation they introduce, but they cannot control how customers perceive it.

More innovation = more users, right?

In non-AI based services, customer behavior follows a clear logic: the more innovative the service is perceived to be, the more satisfied the customer becomes, leading to more users.

Parcel boxes such as Instabox or Budbee are one of many everyday examples. They are new, practical and solve a concrete problem – the result has been both high customer satisfaction and high impact.

With AI, the relationships are different. The study shows that customers are very satisfied with the most advanced AI services, as they relieve time-consuming and complex tasks. AI can help with things that would otherwise take hours – or require specialized knowledge. But when it comes to the willingness to use the services, the researchers are getting an unexpected reaction.

– When the task is important, complex or risky, customers become cautious. They are afraid of handing over too much responsibility to AI,” says Joanna Pilawa.

– A simple AI function that removes an image background is low risk and easy to adopt. But an AI that gives financial advice, while perceived as highly innovative and useful, feels too unsafe for many.

AI must be placed in the right place

The findings mean that companies cannot introduce AI everywhere and expect customers to embrace it. To achieve widespread adoption, AI should be used in peripheral parts of the service, where the customer’s perceived risk is low.

On the other hand, if you want to create high quality for a niche audience, AI can be integrated into the core of the service. This sends a strong signal of innovativeness and leads to high customer satisfaction, but usage is likely to be lower.

– There is no ‘one size fits all’. What matters is that companies understand what part of the service they are changing – and what that change signals to customers.

A new logic for future services

The research results show that companies and developers need to be much more strategic when introducing artificial intelligence into their services. Customers react differently to AI than to previous technological innovations – and this opens up entirely new questions around risk, sense of security and trust.

– AI is a promising technology, but only for companies that understand its paradox: the more innovative the technology, the more the customer needs to feel safe. If companies listen more to how customers perceive innovations, they can both create better services and strengthen their position in the market,” says Joanna Pilawa.

Contact joanna.pilawa@liu.se

More about the thesis
Joanna Pilawa recently defended her thesis Fables for Robots: Service Innovativeness and AI at Linköping University.