Chatbots, disclosure and customer satisfaction
Customer satisfaction and customer engagement are among the fastest-growing keywords in the management journals, and firms keep replacing service agents with chatbots. The EU AI Act requires telling people when they are talking to an AI. Does disclosure lower satisfaction, and can good service recovery make up for it?
What prompted this
In the management journals, "customer satisfaction" had 17 papers in July–September 2026 against about 6 expected, and "customer engagement" 13 against 8 (see the management dashboard). Since October 2025, 90 papers carry a chatbot keyword (54 in AI journals, 29 in management), and only 7 of them also carry a satisfaction, service-quality, customer-service or service-recovery keyword.
Questions
- Do customers rate the same service lower once they learn it came from a chatbot, and does the gap shrink as they get used to it?
- When a chatbot fails, does handing over to a human repair satisfaction better than an apology from the bot?
- Do disclosure rules change which firms deploy chatbots, and for which tasks?
Data that could help
Customer reviews that mention chatbots (Yelp, TripAdvisor, app stores), firms' chatbot launch dates, and field experiments with service providers.
Next step
Collect app-store reviews of banking apps that mention a chatbot, and compare ratings before and after each app began telling users that its assistant is an AI.