📈 economics

Artificial Intelligence in Online Shopping and its Impact on Customer Loyalty: An Integration of TAM and TPB

This study empirically validates an integrated TAM-TPB framework using data from 512 Indian online shoppers to demonstrate that AI perceptions, particularly intelligence and personalization, drive customer loyalty through the sequential mediation of perceived usefulness, attitude toward AI shopping, and purchase intention.

Original authors: Doddahulugappa Goutam, Shirshendu Ganguli

Published 2026-07-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Doddahulugappa Goutam, Shirshendu Ganguli

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you walk into a massive, chaotic supermarket where the aisles stretch for miles, the shelves are stacked with millions of products, and the staff are invisible. This is online shopping. Now, imagine a super-smart, invisible assistant named "AI" that follows you around, knows exactly what you like, and helps you find things instantly.

This research paper asks a simple but crucial question: Does having this smart assistant make you a loyal customer who keeps coming back, or does it just make you buy once and leave?

The authors, studying shoppers in India, built a "recipe" to understand how this works. They combined two famous psychological theories (TAM and TPB) to create a map of the customer's mind. Here is the breakdown of their findings using simple analogies.

The Four Ingredients (The AI Features)

The researchers found that four specific "flavors" of AI experience determine how useful a shopper thinks the system is:

Perceived Intelligence (The "Brain"): This is how smart the AI seems. Does it understand your context? Does it learn your preferences?
- Analogy: It's the difference between a vending machine that only gives you a soda when you press "A1" and a personal chef who knows you hate cilantro and suggests a dish you'll actually love. The smarter the AI seems, the more useful it feels.
Personalisation Quality (The "Tailor"): This is how well the AI customizes the experience for you.
- Analogy: Imagine walking into a store where the mannequins are dressed exactly in your size and style, and the music is your favorite genre. If the AI gets your recommendations right, it feels like a custom suit; if it gets it wrong, it feels like wearing someone else's clothes.
Interactivity (The "Conversation"): This is how natural and responsive the chat feels.
- Analogy: Is the AI like a broken robot that only says "Error," or is it like a friendly shopkeeper who listens, nods, and answers your questions in real-time?
AI-Driven Trust (The "Handshake"): This is whether you believe the AI is honest and safe with your data.
- Analogy: It's the feeling you get when you hand your credit card to a stranger. If you trust them, you feel safe. If you don't, you pull your hand back.

The Chain Reaction (How Loyalty is Built)

The paper argues that these four ingredients don't magically create loyalty. Instead, they set off a domino effect (a sequential chain):

Step 1: The "Useful" Feeling (Cognitive)
First, the AI's intelligence and personalisation make the shopper think, "Wow, this saves me time and helps me find what I need." This is Perceived Usefulness.
- Metaphor: You realize the GPS in your car is actually helping you avoid traffic, not just showing you a map.
Step 2: The "Like" Feeling (Attitude)
Once the shopper realizes the AI is useful, they start to like the experience. They feel positive emotions toward using the AI.
- Metaphor: Because the GPS saved you time, you start to enjoy driving with it. You develop a fondness for the tool.
Step 3: The "I Want to Buy" Feeling (Intention)
Because they like the experience, they form a strong intention to buy.
- Metaphor: You decide, "I'm going to use this GPS again for my next trip."
Step 4: The "Loyalty" (The Result)
Finally, this intention turns into Customer Loyalty. You don't just buy once; you keep coming back and telling your friends.
- Metaphor: You become a brand ambassador for that GPS app.

The Key Findings (What the Data Said)

The researchers surveyed 512 online shoppers in India and ran the numbers. Here is what they discovered:

Intelligence is King: The single biggest factor that made people feel the AI was "useful" was how smart the AI seemed. If the AI acts dumb, nothing else matters.
Trust is the Glue: Trust didn't just help people buy; it was the strongest factor in making people like the AI shopping experience. In a market where people worry about data privacy, trust is the foundation of a good attitude.
The "Like" Factor is Crucial: The study found that Attitude (how much you like the AI) is the most powerful predictor of whether you will actually intend to buy. It's not enough for the AI to be useful; you have to enjoy using it.
It's a Chain, Not a Jump: You can't skip steps. The AI needs to be useful first, which makes you like it, which makes you want to buy, which makes you loyal. If you break the chain (e.g., the AI is useful but you don't trust it), loyalty doesn't happen.

The Bottom Line

The paper concludes that for online stores to keep customers, they can't just slap an AI chatbot on their website. They need to build a system that feels intelligent, personalized, conversational, and trustworthy.

If they do this, the customer's brain goes through a specific journey: "This is useful" → "I like this" → "I want to buy" → "I will stay loyal."

The study suggests that in the Indian market (and likely elsewhere), the quality of the AI experience itself is the new loyalty program. It's not about coupons or points anymore; it's about how well the digital assistant makes you feel understood and safe.

Technical Summary: Artificial Intelligence in Online Shopping and its Impact on Customer Loyalty

Problem Statement and Research Gap
The rapid integration of Artificial Intelligence (AI) into online retail—spanning recommendation engines, conversational agents, and predictive analytics—has redefined consumer behavior. However, the literature remains theoretically disjointed regarding the specific mechanisms by which AI usage translates into long-term customer loyalty, particularly in emerging markets like India. Existing studies tend to isolate either technology acceptance (focusing on TAM constructs) or general loyalty antecedents, failing to integrate AI-specific perceptions (intelligence, personalisation, interactivity, trust) into a unified structural model. Furthermore, the sequential cognitive-attitudinal-volitional pathway linking AI perceptions to loyalty via perceived usefulness, attitude, and purchase intention has not been empirically tested. This study addresses these gaps by developing and validating an integrated Technology Acceptance Model (TAM) and Theory of Planned Behaviour (TPB) framework within the Indian e-commerce context.

Methodology
The study employs a deductive, cross-sectional survey design grounded in a post-positivist philosophical stance.

Data Collection: Primary data were gathered via an electronic survey (Google Forms) between February and April 2025. The target population consisted of adults (18+) in metropolitan and non-metropolitan Indian cities who had used AI-based functions (recommendations, chatbots, virtual assistants) at least twice in the preceding three months.
Sampling: A stratified purposive sampling method was used regarding city levels and gender. From 641 initial responses, 512 valid responses were retained for analysis, meeting the requirements for Covariance-Based Structural Equation Modelling (CB-SEM).
Instrument: The study utilized multi-item reflective scales based on validated prior tools, measured on a seven-point Likert scale. The instrument underwent professional evaluation by six scholars and a pilot study (n=48).
Analysis: Data were analyzed using IBM AMOS 26. The analytical strategy followed a two-step protocol: (1) Confirmatory Factor Analysis (CFA) to assess reliability, convergent validity (AVE > 0.50), and discriminant validity (HTMT < 0.85); and (2) Structural Model estimation with bootstrapped mediation analysis (5,000 resamples) to test sequential indirect effects. Common Method Bias was assessed via Harman's single-factor test and a latent factor approach.

Key Results
The structural model demonstrated excellent fit indices (CFI = 0.945, TLI = 0.940, RMSEA = 0.042) and explained 34.3% of the variance in customer loyalty directly, with the integrated model accounting for 68.3% of the variance in the loyalty construct overall.

Antecedents of Perceived Usefulness: Perceived intelligence emerged as the strongest predictor of perceived usefulness ( $\beta$ = 0.217), followed by AI-driven trust and personalisation quality. This suggests that consumers evaluate the functional utility of AI primarily based on its cognitive capability to understand context and learn preferences.
Antecedents of Attitude: Personalisation quality, perceived interactivity, AI-driven trust, and perceived usefulness all significantly influenced attitude toward AI shopping. Notably, AI-driven trust was a robust predictor of attitude, indicating that confidence in system integrity is crucial for forming positive affective dispositions.
Predictors of Intention and Loyalty: Attitude toward AI shopping was identified as the strongest predictor of purchase intention ( $\beta$ = 0.163) and a strong direct predictor of customer loyalty ( $\beta$ = 0.335). Purchase intention also significantly influenced loyalty.
Sequential Mediation: The bootstrapped analysis confirmed the hypothesized sequential mediation chain: AI Antecedents $\rightarrow$ Perceived Usefulness $\rightarrow$ Attitude toward AI Shopping $\rightarrow$ Purchase Intention $\rightarrow$ Customer Loyalty. Specific pathways, such as Perceived Intelligence $\rightarrow$ Perceived Usefulness $\rightarrow$ Attitude $\rightarrow$ Intention $\rightarrow$ Loyalty, were statistically significant. The study validates that both perceived usefulness and attitude act as necessary, sequential mediators rather than isolated factors.

Key Contributions
The paper claims four primary theoretical contributions:

Multidimensional Antecedent Model: It extends the traditional TAM dyad (usefulness/ease of use) by introducing and validating a four-dimensional AI antecedent model (Perceived Intelligence, Personalisation Quality, Perceived Interactivity, and AI-Driven Trust) specific to retail interactions.
Sequential TAM-TPB Architecture: It provides the first empirical test of a sequential mediation architecture in an AI-mediated retail context, resolving ambiguity about the ordering of cognitive (usefulness) and attitudinal (attitude) mediators in loyalty formation.
Pivotal Role of Attitude: The study identifies "Attitude toward AI Shopping" as the pivotal mediator, demonstrating it is more proximal to loyalty formation than perceived usefulness or purchase intention individually.
Contextual Sensitivity: By situating the model in India, the research contributes contextually sensitive knowledge to a literature dominated by Western and East Asian samples, documenting AI-loyalty mechanisms in a high-uncertainty, mobile-first, price-sensitive emerging market.

Significance and Claims
The authors assert that the study offers a validated theoretical foundation for future AI-loyalty research across diverse cultural settings. Practically, the findings suggest that platform developers and marketers must prioritize investment in the cognitive sophistication of AI (to drive perceived usefulness) and the relational quality of interactions (transparency, explainability, and interactivity) to foster positive attitudes. The paper concludes that in AI-mediated contexts, customer loyalty is shaped primarily by cognitive and attitudinal factors derived from the quality of the AI experience, rather than by transactional incentives alone. The study modestly acknowledges limitations regarding its cross-sectional design (preventing causal inference), potential residual common method bias, and the disproportionate representation of urban/educated consumers, suggesting future longitudinal and rural-focused research.

The Four Ingredients (The AI Features)

The Chain Reaction (How Loyalty is Built)

The Key Findings (What the Data Said)

The Bottom Line

Technical Summary: Artificial Intelligence in Online Shopping and its Impact on Customer Loyalty

More like this