The global chatbot market is predicted to grow from NOK 900 million in 2015 to 10 billion in 2024. In Chatbots for Loyalty, leading Norwegian researchers, industry partners, and chatbot developers join forces in an innovation project supported by the Research Council of Norway BIA programme. The goal for the project has been to create the next generation chatbots for the global market.
The project is owned by artificial intelligence (AI) technology provider Anzyz Technologies. The other industry partners are the chatbot developers boost.ai and EdTech Foundry, as well as Nordea Liv, SpareBank 1 SR-Bank and NKI. Nordea Liv and SpareBank 1 SR-Bank use chatbots for customer service. NKI plan for using chatbots as part of their online platform for education. SINTEF and Center for Artificial Intelligence Research at the University of Agder are research partners.
In the project, we have developed knowledge and technology needed to make chatbots key to companies' service offering. In particular, three research challenges have been addressed. First, the project has generated new knowledge on users' needs and motivations for chatbot use. Second, we have developed machine learning technology to enable chatbots to better predict and act on customers’ needs. Third, we have developed service designs where chatbots are given a more prominent role in the companies' offerings. The goal is that users regard chatbots as an important touchpoint in their relationship with the company, and a preferred first point of contact.
In the project, we have worked on the use of artificial intelligence for analysis of text-based dialogue, designed and tried out prototypes for the use of chatbots to support key customer journeys, and established new knowledge on the design and user experience of chatbots. The work on artificial intelligence is based on a Tsetlin machine approach developed by the University of Agder. Here, an important achievement is the use of this approach to enable efficient identification of new or unknown content in textual material. Chatbots and early versions of related solutions have been designed to support key customer journeys in customer service and education. New knowledge on chatbot design and use have been achieved through a ranger of user studies, including dialogue analyses, experiments, and prototype trials with users.
Virkingene og effektene av prosjektet og prosjektresultatene kan synes på flere nivå:
- Prosjektets partnere: Prosjektet har ledet til ny kunnskap og lagt grunnlag for forbedrede produkter og tjenester. Prosjekteier og de fleste bedriftspartnere benytter prosjektresultatene for forbedring av interne prosesser og for å kunne tilby nye og forbedrede produkter og tjenester.
- Prosjektpartnernes kunder og brukere: Prosjektet og prosjektresultatene har, eller vil bli, brukt for å forbedre tjenester som tilbys prosjektpartneres kunder og brukere – innen områdene helse, kundeservice, og utdanning.
- Akademisk forskning: Kunnskap og teknologi fra prosjektet har bidratt til forskningsfronten gjennom forskningspublikasjoner i tidsskrift og konferanser.
- Samfunnsmessige ringvirkninger: Ved å påvirke tjenester innen helse, kundeservice og utdanning, er prosjektet også forventet å ha samfunnsmessige ringvirkninger – i form av mer effektiv tjenesteyting og bedre opplevelser for brukere.
Chatbots powered by artificial intelligence (AI) transform how service organizations interact with their clients. Chatbots allow users to interact with digital services in natural language , and hold great promises for customer service and education. But current chatbot solutions often fail due to strong technology push with insufficient sensitivity to user needs, and a lack of service design that match AI and machine learning capabilities with users' needs and wants.
In Chatbots for Loyalty, we apply a human-centred approach to innovate chatbot services that strengthen customer loyalty with service providers. We will develop AI capabilities for chatbots to predict and act on customer's likely next best actions. The AI capabilities will be incorporated in service processes where the chatbot is seen as a main point of contact for the customer, integrating service provider offerings.
The project results generate value on three level. Project owner Anzyz Technologies will adapt its AI technology to the chatbot domain. Chatbot providers boost.ai and EdTech Foundry apply the project results in their platforms. Service providers Nordea Liv and SpareBank 1 SR-Bank (customer service user cases) and NKI (education use case) apply the chatbots in their service processes. For Nordea Liv and SpareBank 1 SR-Bank this strengthen customer loyalty with their services . For NKI this strengthen student motivation and course completion.
The chatbot services will be developed in an iterative, case-driven approach, with room for exploration in response to user research and case experiences. Project results are validated and refined in business-critical cases. Key research challenges to be resolved are:
1) Run and benefit from a human-centred approach to chatbot design
2) Customer segmentation and predictions of next best actions in chatbot service processes, based on machine learning from previous chatbot interactions.
3) Service design where chatbots are integrators of service experience and value offerings.