📈 economics

When AI personalisation becomes persuasive: the roles of trust and consumer response in digital advertising

This study utilizes a multi-theoretical framework and survey data to demonstrate that AI-driven personalisation dimensions enhance perceived quality and purchase intention through persuasion, with consumer trust in AI serving as a critical boundary condition that amplifies these effects.

Original authors: My Van Dang, Triet Minh Nguyen

Published 2026-08-03

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: My Van Dang, Triet Minh Nguyen

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 the internet as a giant, bustling marketplace where millions of people are browsing, shopping, and chatting. In this digital bazaar, companies want to show you the perfect item at the perfect moment. To do this, they use a special kind of magic called Artificial Intelligence (AI). Think of AI as a super-smart, invisible librarian who watches what you look at, remembers what you like, and then whispers suggestions into your ear. But here's the tricky part: when this librarian gets really good at guessing what you want, it stops feeling just like "helpful advice" and starts feeling like "persuasion"—a gentle nudge that makes you want to buy something.

This is where trust comes in. Imagine you are walking through a forest. If you trust the guide leading you, you are happy to follow their suggestions, even if the path is a bit mysterious. If you don't trust them, you might stop and question every turn. This study explores that exact feeling in the digital world. It asks: When an AI gives you a perfect recommendation, does it actually make you buy things? And does your trust in the AI act like a supercharger that makes that recommendation even more powerful? The researchers are trying to figure out how these invisible digital nudges turn into real-world shopping carts, and why believing in the AI matters so much.

The Invisible Salesperson and the Trust Switch

So, what did the researchers actually do? They looked at 401 people who had recently seen personalized ads on apps like Facebook, TikTok, and Shopee. They wanted to see how five specific "ingredients" of a good AI ad—Relevance (is it about what I like?), Accuracy (is it right?), Consumer Control (can I say no?), Timeliness (is it showing up now?), and Contextual Fit (does it make sense for where I am?)—work together to make an ad feel high-quality.

They found that all five ingredients matter, but two are the heavy hitters: Relevance and Contextual Fit. It's like a chef cooking a meal; you need the right ingredients, but if the meal is served at the wrong time or in the wrong setting, it just doesn't taste as good. When the AI gets these right, people feel the ad is "high quality."

But here is the real magic trick the paper discovered: High-quality ads don't just make people feel good; they make people persuaded. The study found that when an ad feels perfectly tailored, it doesn't just sit there; it actively convinces you. And this is where the "Trust Switch" flips.

The researchers discovered that trust acts like a volume knob for persuasion.

Low Trust: If you don't trust the AI, even a really good, persuasive ad might just make you suspicious. You might think, "They are trying too hard," and you won't buy.
High Trust: If you do trust the AI, that same persuasive ad becomes incredibly powerful. The study showed that trust significantly strengthens the link between being persuaded and actually buying the product. It's as if trust gives the AI's voice a megaphone. When you trust the system, you are much more likely to listen to its persuasive whisper and reach for your wallet.

What the Numbers Say

The team used a fancy math tool called Partial Least Squares Structural Equation Modelling (PLS-SEM) to crunch the numbers from their survey of 401 people. They didn't just guess; they measured the strength of these relationships.

Here is what the data revealed:

Relevance and Contextual Fit were the strongest drivers of feeling like an ad was high-quality.
Perceived Personalisation Quality had a massive effect on Persuasion (a score of 0.763, which is huge in this kind of study). This means if the ad feels tailored, you are definitely being swayed.
Persuasion then led to Purchase Intention.
Most importantly, the interaction between Trust and Persuasion had a very strong effect (0.399). This confirms that trust isn't just a nice-to-have; it's a critical amplifier. Without trust, the persuasive power of the AI is dampened. With trust, it soars.

The study also found that persuasion acts as a middleman. It's not just that a good ad makes you buy; it's that a good ad makes you feel persuaded, and that feeling is what makes you buy. The ad has to pass through your brain's "persuasion filter" first.

What This Means (and What It Doesn't)

The paper suggests that AI advertising is evolving. It's no longer just a tool for finding information; it's becoming a "persuasive communication mechanism." The authors are careful to say they didn't measure "manipulation" or "loss of freedom" directly. They didn't prove that AI is stealing your will. Instead, they found that when AI is trusted, it becomes very effective at guiding your choices.

They also ruled out the idea that personalization works the same way for everyone. The results suggest that if you don't trust the AI, the whole system is less effective. It's not just about having the right data; it's about having the right relationship with the user.

The Takeaway for a Curious Mind

Think of this study as a map of the digital sales floor. It tells us that companies can't just rely on having the best data (accuracy) or the best timing (timeliness). They have to build trust. If the AI feels like a creepy stalker, the best ad in the world won't work. But if the AI feels like a helpful, honest friend, its suggestions become incredibly powerful.

The researchers admit their study has limits. They looked at young, tech-savvy people in one specific market, so we don't know if this works exactly the same way for older people or in different cultures. They also used a survey, which means they asked people what they would do, rather than watching what they actually did in a store. But the patterns they found are strong and clear: In the world of AI ads, trust is the secret sauce that turns a good recommendation into a "yes, I'll take it."

Technical Summary: When AI Personalisation Becomes Persuasive

Problem Statement and Research Gap
While existing literature establishes that AI-driven personalisation enhances advertising relevance and effectiveness, it largely treats personalisation as a technical targeting capability rather than a psychological driver of persuasion. Current models often overlook the "black box" of how perceived personalisation quality translates into persuasive influence and subsequent behavioural responses. Furthermore, there is limited understanding of the boundary conditions under which AI-mediated persuasion succeeds or fails, specifically regarding the role of consumer trust in opaque algorithmic systems. This study addresses these gaps by investigating how AI-driven personalisation dimensions shape perceived personalisation quality, persuasion, and purchase intention, and how trust in AI advertising moderates these relationships.

Theoretical Framework
The study integrates four theoretical lenses:

Stimulus–Organism–Response (S–O–R) Model: Positions AI personalisation dimensions (relevance, accuracy, control, timeliness, contextual fit) as environmental stimuli that influence internal cognitive states (perceived quality, persuasion) and behavioural responses (purchase intention).
Elaboration Likelihood Model (ELM): Suggests that high-quality personalisation increases cognitive elaboration, making consumers more receptive to persuasive messages via the central route.
Trust Theory: Proposes that trust reduces uncertainty regarding opaque algorithms, acting as a boundary condition that strengthens the link between persuasion and behavioural intention.
Persuasion Knowledge Model (PKM): Used conceptually to interpret how consumers navigate algorithmic targeting, though not directly operationalized as a construct in the empirical model.

Methodology

Design: Quantitative, cross-sectional survey design.
Sample: 401 valid responses from consumers with prior experience interacting with AI-generated personalized advertisements on digital and social media platforms (e.g., Facebook, TikTok, Shopee, Lazada). The sample is predominantly young (18–34 years) and digitally savvy.
Measures: Constructs were measured using adapted, validated scales from prior literature (e.g., Bleier & Eisenbeiss, 2015; Shin, 2021) on a 5-point Likert scale. Key variables include five personalisation dimensions, perceived personalisation quality (PPQ), persuasion, trust in AI advertising, and purchase intention.
Analysis: Partial Least Squares Structural Equation Modelling (PLS-SEM) using SmartPLS software. The analysis included assessment of the measurement model (reliability, convergent and discriminant validity via HTMT) and the structural model (path coefficients, bootstrapping with 5,000 resamples, mediation, and moderation effects).

Key Results

Drivers of Perceived Personalisation Quality (PPQ): All five personalisation dimensions significantly positively influence PPQ. Relevance ( $\beta=0.312$ ) and Contextual Fit ( $\beta=0.287$ ) emerged as the strongest predictors, followed by Accuracy, Consumer Control, and Timeliness.
Impact on Persuasion and Purchase Intention:
- PPQ has a very strong direct effect on Persuasion ( $\beta=0.763$ ).
- PPQ also has a significant direct effect on Purchase Intention ( $\beta=0.464$ ).
- Persuasion significantly influences Purchase Intention ( $\beta=0.233$ ).
Mediation: Persuasion partially mediates the relationship between Perceived Personalisation Quality and Purchase Intention, confirming that personalisation quality must be cognitively processed as persuasive to fully drive behavioural intention.
Moderation by Trust: Trust in AI advertising significantly moderates the relationship between Persuasion and Purchase Intention. The interaction effect is strong ( $\beta=0.399$ ), indicating that the positive impact of persuasion on purchase intention is significantly amplified when consumers trust the AI system.
Model Explanatory Power: The model explains 58.2% of the variance in Persuasion and 72.6% of the variance in Purchase Intention.

Key Contributions and Significance
The authors claim three primary contributions to the literature:

Reconceptualization of Personalisation: The study shifts the focus from personalisation as a mere technological recommendation capability to Perceived Personalisation Quality as a multidimensional consumer evaluation process that activates persuasive processing.
Mechanism of Persuasion: It extends persuasion literature by identifying persuasion as a central psychological translation mechanism. The findings suggest that AI personalisation increasingly functions not just as informational support, but as an adaptive persuasive communication architecture that shapes cognitive evaluations and behavioural responses.
Role of Trust as a Boundary Condition: The study positions trust not merely as an enabler of technology adoption, but as a legitimizing mechanism for algorithmic influence. High trust amplifies the effectiveness of persuasive messages, suggesting that consumers are more willing to accept opaque algorithmic influence when they perceive the system as credible and reliable.

Limitations and Future Directions
The authors explicitly note several limitations:

Construct Scope: The study does not directly measure negative constructs such as perceived manipulation, loss of autonomy, or intrusiveness; these are discussed as conceptual implications rather than empirical findings.
Sample Generalizability: The sample consists of young, digitally savvy users in a single emerging market, limiting generalizability to older demographics or different cultural contexts.
Cross-Sectional Design: The study cannot establish temporal dynamics or causality regarding how trust and persuasion co-evolve over time.
Self-Reported Data: Reliance on stated intentions rather than actual behavioural data (e.g., click-through or purchase logs).

The paper concludes that while AI personalisation effectively drives persuasion and purchase intention, its future development requires balancing persuasive capability with ethical governance, transparency, and the maintenance of consumer trust to mitigate potential tensions regarding autonomy and manipulation.

The Invisible Salesperson and the Trust Switch

What the Numbers Say

What This Means (and What It Doesn't)

The Takeaway for a Curious Mind

Technical Summary: When AI Personalisation Becomes Persuasive

More like this