The Influence of Artificial Intelligence Chatbot Problem Solving on Customers’ Continued Usage Intention in e-Commerce Platforms: An Expectation-Confirmation Model Approach

Studies focusing on intention usually follow extended TAM/ECT/TPB logic, in which quality/usefulness and confirmation lead to higher satisfaction and/or attitudes, which then predict continued use, purchase intention, or loyalty.

“Studies that focus on intention usually follow the extended TAM/ECT/TPB logic, which says that quality/usefulness and confirmation lead to higher satisfaction and/or attitudes, which then predict outcomes like continued use, purchase intention, or loyalty”

Systems designed to minimize friction augment perceived utility and user satisfaction, thereby reinforcing trust and subsequent intention outcomes.

“Systems designed to minimize friction (e.g., expedited information access, simplified input, structured guidance, enhanced navigation) augment perceived utility and user satisfaction, thereby reinforcing trust and subsequent intention outcomes (e.g., retention, purchase intention, reliance-oriented metrics) [6,44,45,48,50,52,67,69].”

Problem-solving proficiency, confirmation, and perceived ease of use increase user satisfaction and trust, which further mediate chatbot usage.

“Gao et al. [48] show that problem-solving proficiency, confirmation, and perceived ease of use increase user satisfaction and trust, which further mediate chatbot usage.”

The net-positive pattern is evident in dominant satisfaction/continuance pathways according to numerous studies.

“According to numerous studies [44,46,48,50,69], this pattern is evident in dominant satisfaction/continuance pathways.”

In the main empirical cluster (Autonomy-1, recommendation/support scope), positive outcomes usually follow a utility/assurance chain: performance cues and interaction quality make people feel more useful and satisfied, which builds trust and leads to downstream adoption/continuance or purchase-intent outcomes.

“In the main empirical cluster (Autonomy-1, recommendation/support scope), positive outcomes usually follow a utility/assurance chain: performance cues and interaction quality make people feel more useful and satisfied, which builds trust and leads to downstream adoption/continuance or purchase-intent outcomes (e.g., [44,46,48,50,69]).”

Perceived usefulness/satisfaction mediating design cues to trust/continuance and trust mediating performance cues to intention are mediation structures frequently reported in the mapped studies.

“Perceived usefulness/satisfaction mediating design cues → trust/continuance and trust mediating performance cues → intention are two examples of mediation structures frequently reported in the mapped studies [44,46,48,50,69].”

The utility/assurance lane predominates in cross-sectional SEM studies within the corpus.

“This lane predominates in crosssectional SEM studies within the corpus [44,46,48,50,69].”

Satisfaction rises via problem-solving, confirmation, usefulness, and interaction/service quality.

“Satisfaction rises via problem-solving, confirmation, usefulness, and interaction/service quality.”

Satisfaction/CX/perceived service quality is mostly positive, but design can introduce negative affect (reactance, overload).

“Mostly positive, but design can introduce negative affect (reactance, overload)”

Satisfaction declines when designs trigger reactance or cognitive/affective overload, which depress downstream evaluations.

“It declines when designs trigger reactance or cognitive/affective overload, which depress downstream evaluations.”