Innovating customer service with AI – combining big language models with local data

About innovation sprint #2

With the launch of Chat GPT at the end of 2022, artificial intelligence (AI) in the form of language models became available and easier to experience and understand for a wider audience. AI creates great opportunities for organisations to develop their businesses, not least in connection with customer communication. However, the use of these language models requires a capacity to review output critically to ensure accuracy and relevance. Knowledge of applications and their effects is still low.

Based on corporate customer communication, five tech companies, together with researchers and students from the University of Borås and students from higher vocational education school in Borås, took on the challenge of working together to make the popular, but difficult to understand, accessible and understandable to more people. They explored what it means and how organisations can use AI to streamline customer communications wherever they occur in the customer journey.

Reflections

An important lesson learned is to be able to use internal company information in dialog with language models without sharing sensitive data.  The Innovation Sprint created great opportunities for research on the application of language models to create business value in different customer situations. The emergence of language models creates new possibilities of action for users compared to traditional information systems. Therefore, it was also important to understand how developers deal with these new opportunities for action. We also see a need for new business modelling techniques to describe and analyse the new opportunities for action.