Divergent Effects of Technological Solutions in Services

No abstract available

The thesis is composed by three essays on the effect of technological solutions in service encounters. In the first essay, “Technology Detour in Personal Service Settings: When Promoting Convenience Inadvertently Reduces Customer/FLE interactions and Backfires the Shopping Experience”, we collaborate with a chain of fashion stores in Italy and conduct different field experiments to investigate the effects of the implementation of narrowcasting technology (i.e., real-time dissemination of information enabled by AI and digital screens) in personal service settings (i.e., service encounters in which customers and frontline employees have rich interactions). We investigated the effect of the narrowcasting technology, aimed at promoting service convenience, on customers shopping behavior. The results suggest that, contrary to expectations, customers who are randomly exposed to narrowcasting technology (treatment group) reduce consumer spending and are less happy with their experience compared to those who were not exposed (control group). The negative effect of the technology is due to the fact that by freeing up customers in their consumption experience, unexpectedly reduces the frontline employee/customer interactions, which is key for the customer experience in the personal service settings. To assess the causality of our mediation model we use moderation of process approach testing the effect of the narrowcasting technology in a situation in which customers are under time pressure (i.e., shopping during lunch break) and the interaction with frontline employees is not pivotal to customer experience. The findings show that during lunch break the narrowcasting technology regain the positive effect that has been found in online and self-service settings by previous scholars. In the second essay, “Using Text Analytics Tools to Improve Online Reputation: Evidence from a Field Experiment and a Lab Study”, we investigate the causal effect of text analytics solutions on hotels’ online reputation through a randomized field experiments: half of hotels in the control group randomly received a report with all their TripAdvisor reviews, in terms of score and raw text, while the other half in the treatment group received the same material plus a document with a sentiment analysis and a topic word-cloud derived from the online reviews. After six months in which such hotels worked to improve their activities, we found the hotels in the treatment group increased their TripAdvisor reputation score 4.2% more than those in the control group, because they were able to identify and act better on the weakest attribute of their offer. The third essay, “The Effects of Artificial Intelligence on Productivity and Quality: A Field Experiment in Banking Services” is in collaboration with a global bank from which we obtained a unique dataset. The aim of this research is to investigate the effect of AI solution implementation om service productivity and service quality. In particular, the bank decided to implement an AI solution aiming to automate the responses to tickets opened by employees who, in its various branches, do not know how to solve issues related to mortgages. The bank accepted to launch the AI solution ‘randomly’ for two months before the full rollout: for these months, a group of branches started to operate assisted by the AI solution, while another group continued to work with the traditional method. This allowed to generate a randomized trial design during that time window. Our results reveal that the AI solution generates the expected efficiency effects on the time to handle the ticket and the satisfaction of the branch employees involved in the process. However, it also produces positive spillovers on customer satisfaction for the mortgages offering and makes, somewhat unexpectedly, treated branches more prudent/selective in granting mortgages.