Effects of Virtual Assistant Features on User Experience in Service Settings

Summary

This study examines how the characteristics and performance of virtual assistants influence user experience in service settings. Approximately 400 adults will interact with virtual assistants presented as avatars, chatbots, or voice assistants in simulated tourism, cultural, retail, or financial service situations.

Participants will be exposed to one or more experimental conditions that may differ in features such as communication style, language specificity, anthropomorphism, personalization, type of virtual assistant, level of immersion, or the presence of inaccurate responses or errors. After the interaction, participants will complete questionnaires about their perceptions, emotions, trust, satisfaction, acceptance of recommendations, and intention to use the virtual assistant.

During the experimental session, non-invasive methods may be used to measure visual attention and psychophysiological responses. These methods include eye tracking, electrodermal activity, electroencephalography, and automated facial expression analysis. Facial recordings will be used to analyze emotional expressions and will not be used to identify participants.

The study does not involve drugs, biological products, medical treatments, diagnostic procedures, or investigational medical devices. Its purpose is to improve scientific understanding of human interaction with artificial intelligence and virtual assistants in service environments.

Conditions

• User Experience With Virtual Assistants in Service Settings

Interventions

BEHAVIORAL

Inaccurate Virtual Assistant Response

The virtual assistant will provide a predefined inaccurate or contextually inappropriate response during the simulated service interaction. The error will be standardized across participants assigned to this condition and will not concern medical, safety-related, or other high-risk information.

BEHAVIORAL

High Anthropomorphism

The virtual assistant will display a comparatively high level of human-like visual, vocal, or conversational characteristics, as operationalized in the experimental protocol.

BEHAVIORAL

Low Anthropomorphism

The virtual assistant will display a comparatively low level of human-like visual, vocal, or conversational characteristics, as operationalized in the experimental protocol.

BEHAVIORAL

Accurate Virtual Assistant Response

The virtual assistant will provide accurate, relevant, and contextually appropriate information during the simulated service interaction.

Sponsors & Collaborators

• University of Malaga

  lead OTHER

Study Design

Allocation

RANDOMIZED

Purpose

BASIC_SCIENCE

Masking

SINGLE

Model

PARALLEL

Eligibility

Min Age

18 Years

Sex

ALL

Healthy Volunteers

Yes

Timeline & Regulatory

Start

2026-09-30

Primary Completion

2028-01-31

Completion

2028-01-31

Countries

• Spain

Study Locations