Effects of Virtual Assistant Features on User Experience in Service Settings (AVir4Services)

Virtual assistants are artificial intelligence systems that enable users to interact with digital services via text, voice, or embodied avatars. Their use does not necessarily improve user experience because interaction outcomes may depend on the design and behavior of the virtual assistant, the service context, and the characteristics of the user.

The main objective of this study is to identify the factors that influence user interaction and user experience outcomes when people engage with virtual assistants in service settings. The study will examine characteristics such as the type of virtual assistant, communication style, language specificity, anthropomorphism, personalization, adaptability, appearance, voice, level of immersion, and the occurrence of inaccurate, erroneous, or unexpected responses.

Approximately 400 adults aged 18 years or older will participate in experimental sessions conducted primarily at the Consumer Behavior Research Laboratory (LICCO) of the University of Malaga. Participants will be recruited through the Online Recruitment System for Economic Experiments (ORSEE), university communication channels, social media, collaborating organizations, and other appropriate recruitment channels.

Participants will interact with or be exposed to virtual assistants developed or configured for the study. These may include avatars, chatbots, and voice assistants presented in two-dimensional or immersive three-dimensional environments. The experimental materials may also include edited images, audio, or video representing simulated interactions with virtual assistants in tourism, cultural, retail, financial, or other service contexts.

Depending on the specific experimental design, participants will be assigned to conditions that differ in one or more virtual assistant features or interaction outcomes. The study may compare different levels of anthropomorphism, communication styles, degrees of language specificity, personalization, assistant modality, immersion, or response accuracy. The final allocation procedures and experimental conditions will be defined before recruitment and documented in the study protocol and ClinicalTrials.gov record.

After the interaction, participants will complete questionnaires assessing variables such as perceived usefulness, enjoyment, symbolic benefits, trust, satisfaction, attitude toward the recommendation, acceptance of recommendations, perceived risk, privacy concerns, emotional response, and intention to use the virtual assistant. Relevant participant characteristics, including age, gender, personality, personal innovativeness, digital experience, and accessibility needs, may also be examined as explanatory or moderating variables.

Non-invasive consumer neuroscience and psychophysiological techniques may be used during the experimental session. Eye tracking will record gaze position, visual fixations, and eye movements. Electrodermal activity will be used to assess changes in physiological arousal. Electroencephalography will record electrical brain activity associated with cognitive and emotional processing. Automated facial expression analysis will code observable facial movements and emotional expressions. Facial recordings will not be used for biometric identification, authentication, or identity verification.

Participants will receive study information and provide informed consent before data collection. A safety screening questionnaire will be administered when required for the psychophysiological procedures. Data will be coded and stored separately from directly identifying information. Results will be reported only in aggregated, anonymized, or appropriately pseudonymized form.

The study is behavioral and non-therapeutic. It does not involve medication, biological samples, medical treatment, diagnosis, or evaluation of a medical device. The procedures are non-invasive, although participants may experience temporary tiredness or minor discomfort during the experimental session.