Augmented Reality (AR) technology combines real images/environments, captured by the camera of PC, mobile phones or other smart devices, with virtual 3D digital graphics/objects, making them coexist in the same space at the same time (Azuma, 1997; Fox, 2012; Kounavis et al., 2012; Rese et al., 2017; Russo et al., 2022). It can be implemented on cutting-edge devices such as Apple Vision Pro or Spacetop AR laptop, as well as on mobile apps for general purpose off-the-shelf smartphones (Arena et al., 2022). Reduced costs and technological advances are expanding AR globally: worldwide revenues are expected to surpass US$198 billion in 2025, and the number of users to exceed US$1.33 billion (Rejeb et al., 2023). According to research data from the IDC Worldwide Augmented and Virtual Reality Spending Guide, the consumer sector accounted for 53% of the market share of AR-related spending in 2020 (Hsu et al., 2021).

This technology digitally reshapes the real world by providing enhanced experience in terms of perceived hedonistic, informative and utilitarian values (Voicu et al., 2023) and it is becoming an attractive means of commercialization and revenue generation (Khan et al., 2015). In fact, it has been estimated that it will be adopted by 59.1% of companies worldwide by 2027 (Dhianita and Rufaidah, 2024).

AR applications span various fields, including healthcare (Douglas et al., 2017; Fida et al., 2018), manufacturing (Eswaran and Bahubalendruni, 2022) and maintenance (Lamberti et al., 2014), ergonomics and work safety (Karlsen et al., 2022), education (Hidayat and Wardat, 2024), vocational training (Chiang et al., 2022) and support to learning disabilities (Shaaban and Mohamed, 2024). They also impact marketing and communication (Rauschnabel et al., 2024) within E-commerce (Shah, 2024) and retail settings (Attri et al., 2024). Notable examples include mobile apps for furniture positioning and layout planning (Alves and Luís Reis, 2020), nutritional information and food tractability (Penco et al., 2021), immersive digital storytelling (Farshid et al., 2018) and pre-purchase evaluation of products (Barta et al., 2023a), especially in the cosmetic sector, with a reported effectiveness in boosting the purchase intentions (Ebrahimabad et al., 2024) and the users’ self-perception and satisfaction (Lee and Oh, 2018; Kosmala et al., 2019).

Psychometric analyses showed satisfactory Cronbach’s alphas in all the assessed variables, ranging between 0.71 and 0.98, meeting the criterion of exceeding 0.70 (Nunnally and Bernstein, 1994).

On average, PEOU and PE exceeded the central point of the scale, with PEOU showing a higher value, while BI did not. Additionally, PD and MI surpassed the central point, while TRUST and DOUBT did not, although TRUST was very close.

High, positive significant correlations were observed among almost all assessed variables, with the strongest correlations found between PI and: PE (r = 0.84); BI (r = 0.78); PEOU (r = 0.72); PD (r = 0.66); TRUST (r = 0.60); and MI (r = 0.39). Furthermore, strong correlations were found between PE and: PEOU (r = 0.81); BI (r = 0.75); TRUST (r = 0.55); and MI (r = 0.43). PE showed a positive but weak correlation with DOUBT (r = 0.19). PEOU also showed positive and relatively strong correlations with other variables, particularly with: BI (r = 0.55); TRUST (r = 0.41); and MI (r = 0.37). Additionally, the correlation between PEOU and DOUBT was weak and positive (r = 0.21).

The following Table 1 summarises the correlations and descriptive statistics for each measure.

The estimated SEM, summarised in Figure 2, showed satisfactory indices of fit, confirming the goodness of the model: χ ² (0) = 0.00 p < 0.00, CFI = 1, TLI = 1, RMSEA = 0.00; SRMR = 0.00. PI showed a direct and positive strong association with PE (β = 0.54), BI (β = 0.50) and TRUST (β = 0.48), and a lower but significant positive association with PD (β = 0.35). PI showed a negative and significant association with DOUBT (β = −0.30), and a non-significant association with MI. PEOU showed a direct and positive strong association with PE (β = 0.41), and a lower but significant positive association with DOUBT (β = 0.35). PEOU showed a negative and significant association with TRUST (β = −0.16), and a non-significant association with BI, PD, and MI. PE showed positive significant associations with all the considered variables: PD (β = 0.38) MI (β = 0.31); BI (β = 0.29); DOUBT (β = 0.29), and TRUST (β = 0.35).

The model supported the mediating role of PE, with all the hypothesized indirect effects significant, as shown in the following Table 2.

Augmented Reality (AR) has been shown effective in providing enhanced product information, addressing the primary limitation of e-commerce in the beauty and cosmetic market, namely, the lack of a pre-purchase product evaluation. This technology has been suggested to create a win-win situation for both companies, which can promote their products online and increase sales, and consumers, who can more easily evaluate products and make informed decisions with the aid of Virtual Try-On (VTO) apps. However, AR technology is still emerging, and more work is required to overcome the psychological challenges that hinder its widespread adoption: the impact of the individual characteristics on consumers’ perceptions and experiences, as well as the influence of the enjoyment on the overall experience, the decision-making processes and the app usage.

Previous studies focused on a limited set of outcomes, namely, the attitude towards AR, the satisfaction with the overall experience and the behavioural intention to use AR-based apps, and did not take into account the complex interconnection of other psychological variables.

To address these limitations, we performed a path analysis based on a Structural Equation Model (SEM), considering 4 dimensions of the Technology Acceptance Model (TAM): perceived informativeness (PI), perceived ease of use (PEOU), perceived enjoyment (PE), and behavioral intention (BI). We additionally included, as outcomes of the model, makeup involvement (MI), perceived diagnosticity (PD), TRUST, and DOUBT.

The observed value of PEOU, PE, and MI, all above the midpoint scale, suggests that VTO cosmetic apps are perceived as simple, user-friendly, and engaging.

The strong correlations between PI and the PEOU, PE, BI, PD and TRUST dimensions suggest its key role. If consumers feel that the app is able to convey information about the products, their perception of user-friendliness increases, as does their enjoyment of using the app itself. Moreover, as long as they feel that they have received detailed information about the products, they are more inclined to purchase the products, which is consistent with the literature (Pai and Yeh, 2014). Additionally, they consider AR as effective in conveying product details, helping them form their opinions, consistent with the literature (Jiang and Benbasat, 2007; Chrimes et al., 2022; Yoo, 2020). PI increases alongside TRUST in the technology and product, highlighting how useful detailed information is in helping consumers feel confident about the product. PI correlation with MI suggests that the more consumers believe they have received interesting information about the cosmetics, the more they tend to increase their relevance and attention to them. Similar results can be found in past studies (e.g., Kim et al., 2010).

The strong correlation between PEOU and PE, also confirmed in previous studies (Holdack et al., 2022; Huang and Liao, 2015), suggests that as much as consumers appreciate the simplicity of the app, they also enjoy the experience and have fun. PEOU also correlates positively with BI, which is in line with other studies on mobile wallets (To and Trinh, 2021) or on e-business environments (Leyton et al., 2015). Another strong correlation is with PD, confirming that the easier an app is to use, the more you feel that it can provide all the information you need to form a correct opinion about the product. However, it is interesting to note that there is a low significant but positive correlation with DOUBT, leading to the suggestion that perhaps excessive ease of use may raise doubt rather than convey confidence in the product. This ambiguous role of PEOU is sometimes found in the literature, especially in areas where security is an issue (Al-Sharafi et al., 2016), suggesting that something too easy can be interpreted as risky in the perception of the users.

PE shows a high correlation with BI and PD, indicating that enjoyment in using the technology increases alongside the perception of its effectiveness in explaining product details, making enjoyable experiences more likely to lead to purchase intentions. PE also shows a correlation with TRUST, that is, people who feel happy and engaged when using beauty apps are more inclined to be confident about the technology and the product. This has also been found in other studies on the use of online banking systems (Rouibah et al., 2016), mobile wallets (To and Trinh, 2021), and online shopping (Marza et al., 2019; Gao and Wu, 2010). Finally, PE also correlates with MI, suggesting that enjoyment and engagement in makeup use tend to grow side by side, consistently with other studies (e.g., Shiau and Luo, 2013).

BI correlates with PI and PE, in line with much of the literature in this area (e.g., Oyman et al., 2022), and shows that the intention to purchase products presented via the AR app is related to how much people believe they will receive adequate information about the products themselves (Hameed et al., 2024), and how much people enjoy using the technology (Ramayah and Ignatius, 2005; Bashir and Madhavaiah, 2015). BI also correlates with PEOU and although the relationship between these two variables provides two-faced results (Li and Chu, 2024; Kim and Kang, 2012), some literature still confirms that purchase intention tends to increase with the perception of ease of use (Ramayah and Ignatius, 2005).

PD correlates very strongly with PI, PEOU, PE and BI and interestingly also with TRUST. Therefore, our ability to evaluate and understand the quality of products increases with the perceived ease of use and enjoyment of the app, as well as with our purchase intention (Cheng et al., 2022) and trust in the product (Avnet et al., 2012).

TRUST shows strong correlations with PI, PD and BI, indicating that the informativeness about the product allows users to have confidence in it and to be sure that the AR app provides the appropriate information that allows to give a clear idea of how to make the purchase (Alalwan et al., 2018).

DOUBT shows only a mild yet significant positive correlation with PEOU and PE, indicating that it is not simply the inverse of TRUST but a distinct dimension with different behavior, despite both being key factors for technology acceptance (e.g., Rese et al., 2017). For a cosmetics app to be fully embraced, it should strike a balance - neither too easy to use nor overly playful - since consumers may become suspicious if the experience feels too simple or too entertaining.

The path analysis reveals key findings, starting from the strong positive associations between PI and TRUST, BI and PD. This suggests that obtaining product information influences trust, purchase intention, and the clarity of consumer perceptions about the product. Therefore, H1a for PE, BI, PD, and TRUST is confirmed.

PI also shows a negative association with DOUBT, indicating that clear, detailed product information reduces consumer doubts about whether the product will meet expectations, confirming H1b.

PEOU shows a positive association with PE, confirming that ease of use leads to enjoyment and relaxation. However, PEOU is negatively associated with TRUST and positively associated with DOUBT. This supports the idea that if something is too easy to use, it can cause skepticism. Therefore, H2a is confirmed only for PE, while H2b is not.

PE correlates positively with BI, PD, TRUST, and MI. This suggests that enjoyment while using the app positively impacts purchase intention, product trustworthiness, and engagement with the product, fully supporting hypothesis H3a.

However, PE also correlates positively with DOUBT, indicating that if the app is too playful, consumers may perceive it more as a game than a shopping tool, increasing doubts about the purchase process. Thus, H3b is rejected.

Regarding the mediator role of PE, PI has both direct and indirect effects (via PE) on BI, PD, and TRUST, but no direct effect on MI. This contrasts with previous findings (Kim et al., 2010), suggesting that clear product information alone does not lead to product involvement. However, when information is conveyed in a playful manner, it does increase consumer involvement. Additionally, information reduces doubt, but when presented in an enjoyable way, it can increase skepticism. This is consistent with the correlation results and highlights that doubts arise when an app feels too playful. Therefore, PE mediates the relationship between PI and BI, PD, MI, and TRUST, confirming H4a. The mediation of PE between PI and DOUBT is positive, emphasizing that enjoyment should be balanced.

As for PEOU, it has no direct influence on BI, PD, or MI, indicating that ease of use alone does not impact behavioral intention or perceptions about the app’s effectiveness. However, when PE mediates, ease of use positively impacts purchase intention, product quality perception, and involvement with cosmetic products. The playfulness of the app significantly impacts all these outcomes, suggesting that if the app offers a hedonic experience, the pleasure derived from it boosts purchase intention, involvement, and the perception that the app helps form a clear opinion about the product. However, ease of use needs to be carefully measured, as PEOU has a direct negative impact on trust, with users tending to perceive the app as untrustworthy. If pleasure mediates this relationship, trust increases. Therefore, PE mediates the relationship between PEOU and BI, PD, MI, and TRUST, confirming H4b. PEOU also has a direct and indirect positive effect on DOUBT, confirming that doubts about product trustworthiness increase when the app is too simple, supporting earlier correlations. This suggests that for a cosmetics app to be effective and reliable, it should not appear too simple.

This study provides three notable insights. First, PE plays a critical role in AR app usage. Enjoyment has been a consistent driver of attitudes toward AR adoption (Won et al., 2023) and a predictor of purchase intention (Holdack et al., 2022). Although results are sometimes less clear (Iranmanesh et al., 2024), PE remains decisive in the TAM model (Alalwan et al., 2018; Balog and Pribeanu, 2010), contributing to positive evaluations of both the app and the brand, motivating users, and encouraging future use (Kim and Forsythe, 2008). It also frequently acts as a mediator in many TAM-based models (Basari and Dewanti, 2024; Hasan et al., 2021).

The second insight is the role of DOUBT, particularly in relation to PE and PEOU. While enjoyment boosts decision-making, PE shows a dual nature when DOUBT is considered. Correlations confirm that if the app feels too much like a game, consumers will doubt its reliability and may not use it for shopping. The indirect effects, where PI’s negative association with DOUBT turns positive through PE, support this. Although detailed product information helps reduce doubt and positively influences decision-making, a playful interface may undermine the app’s seriousness. Furthermore, DOUBT is significant in its relationship with PEOU, suggesting skepticism towards AR increases when the app is too easy, confirmed by the negative association between PEOU and TRUST. Future research should explore the relationship between TRUST and DOUBT, which was not significant in this study.

The last insight involves PI, particularly its relation to PE. PI is a critical component in TAM-based studies and a reliable predictor of attitudes and purchase intention (Rese et al., 2017; Smink et al., 2019; Kim et al., 2010). It is the only variable that negatively correlates with DOUBT, highlighting its key role in reducing uncertainty and resistance to purchase. In this study, PI serves as a strong cognitive predictor, interacting effectively with all outcomes. PE mediates the relationship between PI and the outcomes, suggesting that product information is crucial for behavioral intention, trust, and PD.

Since this study focused on the user/consumer experience rather than on technical aspects of AR apps, we believe it offers valuable insights for both the cosmetics industry and the academic community.

Previous research has shown that gamification elements in AR applications can increase user enjoyment, especially in public spaces (Noreikis et al., 2019). However, gamification should not be the primary feature of an app, nor should ease of use be the sole focus, as excessive playfulness or simplicity may lead users to question the app’s purpose and increase uncertainty. Cosmetics companies should aim to balance enjoyment and ease of use in app development to avoid user frustration and ensure the app remains engaging and practical (Dirin and Laine, 2018). Continuous attention to user experience is essential to encourage repeat use of the app or platform. The academic community should further explore the role of perceived enjoyment (PE) in AR applications across various sectors to better understand their adaptability, as demonstrated in the tourism industry (Aburub, 2023).

The study also underscores the importance of informativeness in influencing purchase intentions, as shown in previous research (Tabaeeian et al., 2024). Companies should prioritize providing detailed product information in their AR applications to facilitate consumer decision-making. From an academic perspective, the study reinforces the significance of informativeness in B2C contexts and raises the question of whether AR technology could be applied similarly in B2B contexts, such as through interactive presentations that engage potential clients.

While this study presents promising implications, it is important to acknowledge its limitations. Future research should address the non-probabilistic sample used in this study. Although it includes various age clusters, it is slightly skewed toward women and younger individuals, who are the primary cosmetics users (Ramshida and Manikandan, 2014). To validate the findings across different demographics, future studies should include a more uniform sample, in both age and gender variables.

Since it included only Italian consumers, further research examining European and non-European populations is suggested. From a cross-cultural perspective, this research would provide a clearer view of consumer behavior regarding AR technology.

The study did not compare the AR experience with a traditional product test without AR. Such a comparison could offer deeper insights into the added value of AR in consumer decision-making.

Since the data are only based on self-reports, the results may be subject to potential biases, as participants’ perceptions and behaviors might not fully reflect their actual experiences.

Finally, as conducted within the beauty and cosmetics, the findings may not be directly applicable to other market sectors. Thus, there is a need for more comprehensive studies to explore the potential and limitations of AR applications across various contexts.

This study explored how the complex interaction of multiple factors (identified as antecedents, mediators or outcomes) influences the use of cosmetics Virtual Try-On (VTO). Dimensions of the Technology Acceptance Model (TAM) and additional constructs were evaluated through a path analysis based on a Structural Equation Model (SEM).

The results confirm, first, the robustness of TAM in assessing AR-based apps, consistent with the literature (Rese et al., 2017; Oyman et al., 2022). Furthermore, Perceived Informativeness (PI) and Perceived Ease of Use (PEOU) were identified as predictors, while Perceived Enjoyment (PE) acted as a mediator, influencing multiple outcomes related to app usage, product evaluation, and decision-making (Iranmanesh et al., 2024). The interaction of these factors offers a more comprehensive understanding of the consumer experience with AR apps. PI emerged as a strong predictor of Behavioral Intention (BI), highlighting the importance of information in consumer decision-making, while PE was a critical mediator. Furthermore, moderate levels of PE and PEOU were found to enhance TRUST and reduce DOUBT. Both Perceived Diagnosticity (PD) and BI were positively impacted by the AR experience, with their coexistence being essential for effective app usage.

AR provides a multisensory simulation that enhances the shopping and entertainment experience (Huang and Liao, 2015). Therefore, Perceived Enjoyment (PE) is essential for understanding the use of cosmetics apps. Throughout this study, PE has been shown to be a key element of the customer experience in the digital environment. The hedonic experience consumers enjoy can be both pleasant and engaging, complementing the cognitive experience associated with product information. This combination allows consumers to perceive the AR app as useful for understanding the product and making a purchase, while also fostering engagement with the products and enhancing their perception of reliability. Finally, attention must be given to the activation of DOUBT, as it acts as a barrier to both purchase and app usage. Therefore, it is crucial to consider it when developing a cosmetic AR app, to prevent resistance and ensure a smoother user experience.