Integrating sensory marketing with artificial intelligence in hospitality: a future research agenda

Scholars and practitioners have shown a growing interest in sensory marketing (SM) within hospitality contexts (Chen, 2025; Fong et al., 2023; Spence, 2022). SM stems from consumer psychology and explores how sensory experiences influence consumer behaviour (Akarsu et al., 2022; Krishna and Schwarz, 2014). Although SM’s relevance and application is acknowledged in hospitality (Fong et al., 2023), research in this domain remains limited (Aksenova et al., 2022; Lee et al., 2019).

The future of hospitality lies in designing unique and memorable experiences, classified as physical, virtual or hybrid (Zhang et al., 2024), influencing consumers’ behavioural intentions (Shin et al., 2024). However, hospitality research has neglected key SM concepts, such as cross-modal correspondence, sensory (in)congruence and sensory intensity (Fong et al., 2023). Such a gap stems from the difficulty of effectively integrating sensory cues to influence the multi-sensory experience, highlighting the need for future research. However, as SM is multidisciplinary, there are challenges in generalising findings across different contexts (Wörfel et al., 2022). The domain’s diversity, drawing from psychology, marketing and sensory science, complicates the forging of a unified research direction. While SM enriches the domain with various perspectives, such eclecticism makes it challenging to distil pathways for further investigation.

The hospitality sector is being transformed by artificial intelligence (AI) and its subsets, such as machine learning (ML) and sensor technologies (Huang and Rust, 2022) which reshape guest experiences and operations (Mariani et al., 2023; Robinson et al., 2020). However, the lack of comprehensive methods to capture sensory experiences limits understanding across contexts (Eklund and Helmefalk, 2022). Technical advancements now offer new ways to measure and interpret sensory data (Doborjeh et al., 2022). Rapid growth in AI and deep learning enables analysis of complex data sets and reveals patterns beyond human ability.

Despite these advancements, the application of SM aided by AI within hospitality is overlooked, given its potential to enhance consumers’ sensory experience. While AI solutions may assist in uncovering patterns (Filieri et al., 2021), sensory experiences are complex, situational and individual (Akarsu et al., 2022), especially when hospitality settings vary. The fallout is a myriad of potential data points to measure and experimentally gather, which usually results in many smaller and limited samples. Using AI for SM in hospitality would require a deeper understanding of what, how, where and who is measured. Despite potential hurdles in gathering data, research illuminates the growing importance of technology integration and states that further attention is needed to explore the incorporation of sensory cues and AI-enabled technology in retail or hospitality settings (Pandey and Tripathi, 2025). Consequently, hospitality research should evaluate SM when designing multi-sensory hospitality experiences (Agapito, 2020). As algorithms and technology evolve, these developments challenge conventional research methods and may drive theoretical advancements. In this context, suggesting that AI and ML methodologies address existing challenges in SM research and foster theoretical innovation. Thus, the current paper is theoretical, aiming to provide a basis for merging SM with AI and to propose future research avenues for further validation and testing.

A semi-systematic literature review was used to guide such theoretical exploration. Following the Snyder (2019) logic, the purpose of such a review is to provide an overview of a research area with a broad research question. Furthermore, the review explores how research within the given domain has evolved by identifying, analysing and reporting patterns in terms of themes. The contribution of the evaluation lies in synthesising current knowledge in the given field to engender future research avenues. Therefore, the study’s research questions are:

Hence, this study aims to:

• conduct an extensive SM literature review to identify contemporary issues;

• conceptualise the identified issues with AI solutions in hospitality; and

• propose future research avenues.

The intersection of SM and hospitality aided by AI presents ample opportunities for developing innovative strategies to enhance the consumer experience, which are discussed as a result of this study.

SM encompasses researching the influence of sensory cues – such as from visual, auditory, olfactory, haptic and gustatory cues – on sensations, cognition, emotions and behaviours (Krishna, 2012). Sensory cues and their effects are examined across various marketing domains (Pandey and Tripathi, 2025) and contexts. SM is “marketing that engages the consumers’ senses and affects their perception, judgment and behavior” (Krishna, 2012, p. 333), but also influence physiological responses (Ruzeviciute et al., 2019). Sensory experiences are subjective and are essential in influencing consumers’ actions, cognitions, behaviours and emotions (Krishna and Schwarz, 2014). SM stems from environmental psychology, inspired by the stimulus-organism-response (S-O-R) model (Mehrabian and Russell, 1974), which is often used to examine atmospheric and sensory cues in different service settings. For instance, by applying the S-O-R model, research has demonstrated the effect of visual stimulus/cues (e.g. glossiness) on haptic perception (internal response), influencing consumers’ reactions (behavioural intentions) (Briand Decré and Cloonan, 2019).

Building on the idea that experience arises from sensory stimuli, research in tourism and hospitality has examined how specific cues shape affective, cognitive, sensory, social and behavioural components (Agapito and Sigala, 2024; Veloso and Gomez-Suarez, 2023). Customer experience has received substantial attention in hospitality (Kim and So, 2022). Within this framework, sensations – an antecedent of brand experience (Chang and Cheng, 2023) – are integral to the hospitality experience (Agapito and Sigala, 2024). For instance, studies have highlighted the importance of human senses in segmenting rural tourists (Agapito et al., 2014) and underscored the relevance of multi-sensory experiences (Molina-Collado et al., 2024). Although experiential consumption is widely acknowledged in hospitality, key concepts of SM, such as multi-sensory experience, sensations, sensory cues, sensory overload and perception, remain underexplored (Pandey and Tripathi, 2025).

The application of SM to hospitality presents a novel avenue for future research. Focusing on sensory congruence and intensity could deepen understanding and advance SM in the domain (Fong et al., 2023). Such insights would enrich the design of hospitality services and inform value creation strategies (Spence, 2022).

Hospitality research highlights AI’s transformative role in sharing visitor journeys (Filieri et al., 2021), engancing experiences, service environment and offerings (Law et al., 2024). AI personalises and engages customer journeys (Orea-Giner et al., 2022) through sensors, facial recognition and environmental monitoring, improving service standards (Ghesh et al., 2023). Increasingly, hospitality embraces AI-driven SM to anticipate preferences, enriching experiences and influencing visit intentions (Fong et al., 2023). AI algorithms also help analyse customer data, enabling tailored marketing and enhancing experiences on a subconscious level (Ghesh et al., 2023).

We distinguish between learning-based AI (e.g. ML, deep learning and reinforcement learning), which learns from data, and non-learning AI (e.g. rule-based systems, logic-driven expert systems, or traditional automation), which follows fixed rules without adaptation. This distinction is key to understanding their roles, data needs and personalisation potential in SM. ML algorithms, including artificial neural networks, classical methods and reinforcement learning, drive innovation in hospitality (e.g. Herhausen et al., 2024; Russell and Norvig, 2021). Convolutional neural networks (CNN) specialised in image processing, support facial recognition and visual data analysis for personalised services. Natural language processing (NLP) powers chatbots, translation and sentiment analysis, enhancing guest interaction (Piris and Gay, 2021). Generative AI creates tailored content (e.g. text, images or music), supporting personalised marketing (Dwivedi et al., 2023).

Classical learning methods, like supervised learning (training models with labelled data) and unsupervised learning (pattern detection in unlabelled data), support customer segmentation, predictive analytics and demand forecasting. For example, decision trees or logistic regression can predict booking behaviours, while unsupervised methods identify customer groups to targeted promotions (Russell and Norvig, 2021). Reinforcement learning optimizes hotel operations, such as energy use or inventory management based on guest behaviour (Prentice et al., 2020; Ruel et al., 2021).

AI also enhances SM by delivering brand-aligned, immersive sensory experiences that boost satisfaction and loyalty (Gonçalves et al., 2024; Pelet et al., 2021). Its ability to target senses enables innovative SM strategies for personalised customer experiences (Mehta et al., 2022). For an overview of the intersection between hospitality, SM and AI, including practical examples and research, see Tables A1 and B1 in the supplementary file.

A sequential approach was undertaken to address the research questions (Palmatier et al., 2018). First, an extensive literature review was conducted to identify contemporary issues in SM; and, second, the issues were conceptualised with AI and specifically ML solutions in hospitality and management contexts, which generated future research avenues. The review was used in three stages. Such an approach suits a theory-based systematic review (Paul and Criado, 2020) with a semi-systematic nature (Snyder, 2019). The procedure was inspired by and included the seven steps of Palmatier et al. (2018) and the rules of thumbs of Paul and Criado (2020). For details of each stage and steps to ensure validity, see section C1, D.1 and H1 of the supplementary file. (1) The first stage of the review was a semi-systematic literature review with a keyword search in different databases; sensory cues AND sensory marketing OR crossmodal correspondence OR atmospheric* OR congrue*. Inclusion and exclusion of articles was performed with the assistance of labelling and training a system to find relevant articles, with the ML tool, ASReview. It yielded 220 relevant articles from a data set of 4,051 articles with no limit regarding publication year (see Table F1 in supplementary file for journal frequency). The results were manually checked for and several procedures were undertaken to ensure the validity of the dataset. (2) The second stage articles from stage 1 and undertook a thematic analysis. The analysis allows for an in-depth, qualitative exploration of the ideas and arguments (Braun and Clarke, 2006), enabling a rich understanding of the research field (Vaismoradi et al., 2013). As such, a thematic approach (Lim et al., 2022) was used with the assistance of MaxQDA, a type of CAQDA tool. Following conceptual and methodological issues and limitations, the results from first and second order coding, resulted in six overarching themes in SM. The process is illustrated in Tables C2, C3 and Figure E1 in supplementary file. Regarding the trustworthiness of the thematic analysis procedure, we followed similar procedures to Nowell et al. (2017) and Scandura and Williams (2000). To address how AI can enhance SM strategies in hospitality and what specific opportunities arise from integrating AI into hospitality, a conceptual integration between the findings in Stage 2 was needed. (3) Stage 3 used the procedure of Fan et al. (2022), where SM was used as a conceptual tool to integrate and synthesise AI in hospitality contexts. The integrative procedure juxtaposed creative generation and rigour through the constant interchange of ideas, writing, reading and discussing various AI and SM conceptual integrations in hospitality contexts (Fan et al., 2022). Subsequently, future research avenues were proposed, emphasising different solutions and areas of enquiry, elaborated upon in the following section.

The review, coding and findings revealed six themes: atmospheric and sensory cues, (in)congruence, multi-sensory cues, perceptions, sensations and overload, of which issues in SM were categorised accordingly. The findings and common issues in SM are summarised and discussed in relation to the themes with hospitality and AI/ML. Table 1 is developed according to the themes revealed in postulated research propositions for SM in hospitality to summarise future research avenues revealed in this study.

Regarding generic atmospheric and sensory cues in SM research, the findings emphasised the lack of knowledge on the interplay between physical and digital contexts. Cues are measured in various ways, providing challenges, such as their dispersion, contamination and other influencing factors in physical settings (e.g. scents or interfering music). Another identified challenge was establishing cue intensity and congruence to avoid sensory overload. In this vein, AI can learn to optimise the intensity and congruence of sensory cues in various settings. Through predictive analytics, firms can find the “sweet spot” for cue deployment, enhance the sensory experience without overwhelming customers (Lecointre-Erickson et al., 2018), and manage multimodal data, while simultaneously processing and analysing data from various senses. Such capability could be instrumental in determining the optimal combination of sensory cues to enhance the atmosphere without causing sensory overload (Flavián et al., 2021).

Research indicates the challenges that advanced technologies could address (Rodgers et al., 2021), such as bridging the gap between digital and hospitality. By analysing customer data, ML assists in tailoring the shopping environment to match cultural preferences or adapt to mood changes (Loureiro et al., 2021). Advanced technologies, such as AI-driven sensors and Internet of Things (IoT) devices, accurately isolate (e.g. scent diffusers and speakers) and measure the impact of sensory cues in real time. For instance, some hotels are exploring implementing AI-powered systems that use sensors and guest data to personalise the in-room ambiance and atmosphere, including lighting, temperature and scent, to enhance guest comfort and satisfaction (e.g. Yin et al., 2023).

The issues of generalising findings were revealed. AI assists in navigating the vast amount of data required to determine cues for optimal evaluations, behaviours and revisits at different destinations and service settings. Regarding the visitor journey, AI’s advantage is to gather multimodal sensorial data through visits (Filieri et al., 2021) or, for instance, through robots, virtual travel, augmented reality (AR) or chatbots (Doborjeh et al., 2022). Addressing the interaction effects through traditional experiments and other methods is time-consuming. In hospitality, unsupervised learning techniques and reinforcement learning (see Table A1) could further improve these designs for businesses (Cherenkov et al., 2024). For example, hotels can track customer journeys based on sensory preferences, such as the need for touch (NFT) (Peck and Childers, 2003), and identify patterns influencing comfort. Unsupervised learning techniques, including clustering algorithms, can segment consumers based on sensory cue preferences. Moreover, reinforcement learning can dynamically adjust sensory cues in hospitality. For instance, predictive models analyse real-time data (e.g. occupancy, guest profiles) to suggest optimal lighting, music and scents for specific times of day or guest demographics.

Research shows that sensory and non-sensory factors impact satisfaction, retention, and positive word of mouth, which posits that understanding this on a broader level becomes important (Muskat et al., 2024). Furthermore, cue intensity and congruency between sensory cues in an atmosphere impact the pleasure of an experience (Spence et al., 2014). Such considerations are relevant when designing hospitality settings and services, finding optimal multi-sensory cues and offering flexibility in transferring findings across different destinations. Hence, we propose:

Congruency and its different outcomes are explained using different theories. Congruency emphasises similarity or consistency, impacting consumer expectation and confirmation (Eklund and Helmefalk, 2022). Psychological mechanisms explain positive evaluations and behaviours from being subjected to congruent cues, such as salience, fluency or emotions (e.g. Spence, 2022). One methodological issue is often establishing and confirming the internal validity of congruency between two objects. Previously, this was achieved by pretests, surveys or qualitative methods (Spangenberg et al., 2006). Research shows that incongruency between cues or objects causes an increase in cognitive processing or disconfirming of expectations (Huang and Wan, 2019). Furthermore, the personal relevance of congruent cues is essential, suggesting a further enquiry into how AI tailors these combinations for individual profiles (Rodgers et al., 2021), while maintaining brand integrity. To exemplify deconstructing a brand into a semantic network (see Eklund and Helmefalk, 2022), NLP and generative AI can deconstruct brand elements and suggest complementary sensory features (e.g. scents to match visuals), ensuring alignment with cultural contexts or individual preferences (Errajaa et al., 2021). Reinforcement learning can ensure congruency by adapting in real time to diverse consumer groups. Similarly, research posits that technology identifies (in)congruent cues within digital storefronts, enhancing customer interaction through strategic sensory alignment (Lecointre-Erickson et al., 2018). Thus, future research is suggested to investigate how AI will suggest sensory congruence, not just at the individual level, but across different cultures.

The level of congruence determines the efficiency by pairing sensory cues, such as the fit between colour, music and a hotel bar (Lin, 2009) or the scent and the virtual destination (Flavián et al., 2021). A potential solution is to use NLP or generative AI to find elements that match with another as well as guest preferences or profiles (Wang and Uysal, 2024). Research reveals perceived congruency in hospitality settings, such as between the atmosphere, theme, food, exterior and interior décor. Findings show a positive influence on the pleasure level of consumers (Lin and Mattila, 2010). Moreover, the abstract pairing was revealed, such as internal communication in hospitality contexts between employers who have received training in brand management and employees’ congruence perceptions of the brand (Kang et al., 2019). A destination’s sensory properties have been examined by measuring the sensory dimensions of hospitality experiences (Buzova et al., 2021). However, AI assists in obtaining deeper insights, such as automatically developing appropriate music playlists or scents at different restaurants and lobbies, while being consistent with the brand. The concept of congruency is also beneficial for the design of robots for human–robot interaction and adaptation in service and hospitality settings (Ma et al., 2024). Therefore, we postulate:

Multi-sensory cues are primarily examined in settings where consumers use all senses (Wörfel et al., 2022). The limitation of multi-sensory cues is the measurement, and research tends to focus on a few combinations of interactions (Helmefalk and Hultén, 2017). While specific cues, such as visual, scent and music, remain popular in physical settings, research still emphasises the lack of SM in digital contexts calling for innovative approaches to simulate comprehensive sensory interactions. For instance, haptic-feedback-device technology enriches digital experiences, otherwise missed, as suggested in SM (Petit et al., 2019). A critical area for further investigation is applying and integrating AI and virtual reality (VR) technologies (Flavián et al., 2021). Such tools offer deeper insights into the effectiveness of multi-sensory cues across physical and digital settings. One application is AI-powered VR tours that incorporate multi-sensory elements like ambient sounds and even subtle scents to provide potential guests with a more immersive and realistic preview of an experience, influencing their booking decisions (Martins et al., 2017).

In hospitality, multi-sensory cues relate to the design of a setting, such as a hotel, restaurant, or tourist destination. In contrast to place branding that considers broader perspectives regarding destinations (Oliveira, 2015), integrating multi-sensory cues in hospitality is holistic, using the gestalt approach, for instance, in digital museums (Guo et al., 2021). Research emphasises how hotels apply SM to enhance visitors’ experiences, ranging from biophilic design and temperature to food (Huang et al., 2024). Surprisingly, the application of SM in hotels is scarce (Spence, 2022). This provides opportunities using emerging pattern mining to identify hotel preferences, VR and sensescape design.

Regarding the holistic perception of multi-sensory atmospheres, research highlights that the sum is larger than the individual parts (Choi and Kandampully, 2019). Correspondingly, it is emphasised that the challenges of identifying and examining all the interactions in a setting would result in an unmanageable number (Ballantine et al., 2010). Hence, introducing novel technologies and possibilities to obtain data (Pelet et al., 2021) opens up new ways of examining potential cue combinations and their calculated effects.

Practical examples can include using digital twins of hotel rooms or virtual tours of resorts using personalised AI-enabled VR to simulate and test multi-sensory experiences before actual implementation. For instance, virtual tours can include auditory, visual and scent-based cues to offer holistic previews, influencing customer decisions (Law et al., 2024). Hence, we propose:

Perception is an essential facet of SM and considers consumer perception of sensory cues in isolation or combination (Krishna, 2012). Current issues in research reveal the need to understand how to best augment and measure experiences with technology. Several methods, such as inclusive biometric methods, can be used to obtain detailed information. For example, AI-powered facial recognition and emotion analysis tools can provide real-time feedback on guests’ reactions to various sensory cues in a hotel lobby or restaurant, allowing immediate adjustments to enhance their experiences (Huang et al., 2024) and offering an objective view compared to traditional methods for measuring perceptions of product placement and store layouts. One potential solution is for hotels to employ broad NLP or generative AI and deep learning to personalise and enhance guest experiences by understanding and responding to their sensory preferences and feedback. However, a discussion is needed about the best practices for leveraging such technologies in SM research (Javornik et al., 2021). The difficulty lies in enhancing consumer experience through technology and ensuring the findings are generalisable and supported by longitudinal data to strengthen validity.

Perception considers experiences and bias formed by sensory cues. It involves various perceptions that influence other relevant outcomes in hospitality (Chang and Cheng, 2023). Although research shows that bias (heuristics, social bias, stereotype, framing effect and cognitive dissonance) occurs when individuals select hotel destinations (Wattanacharoensil and La-ornual, 2019), visual cognitive bias is less studied, such as the size of shapes and perception of volume. While methods exist to analyse sensory cues in hospitality, such as visual analysis (Lobinger and Mele, 2020), research is encouraged to explore how perceptions of other senses are distorted, enhanced or changed in different contexts. While SM research on perception is abundant, cross-context application remains challenging. Scents and other cues impact cognitive and behavioural responses more in hospitality than in retail. Although hospitality uses biometric data, researchers recommend applying cognitive psychology to better interpret it (Walters et al., 2023). SM helps explain cause-and-effect relationships over time, advancing hospitality research through AI and ML. With technology and data integration in mind, we propose:

Research reveals that sensations experienced subconsciously influence consumer outcomes (Krishna and Schwarz, 2014). Moreover, situations that include time are essential, such as memory fading for reliable measurement, meaning research can become resource-demanding in checking for manipulations. SM still has challenges regarding measuring sensations to overcome. Furthermore, findings reveal that research has focused on positive aspects, while negative sensations like hunger and subconscious processes have been overlooked. In addition, traditional methods rely on self-reporting, which can be biased and not accurately capture the intensity or nuance of the sensory experience. ML algorithms can analyse large data sets of consumer interactions to detect patterns and nuances in sensation responses, even those that are subconsciously influenced. This could lead to a more robust understanding of positive and negative sensations as well as consumers’ well-being. Furthermore, portable technologies (e.g. smartwatches and biometric wearables with sensors) assist in capturing the temporal aspects of sensations (Petit et al., 2019), such as how sensory experiences fade over time, which is crucial for measuring and understanding the longevity of SM effects. Simultaneously, a discussion of ethics and integrity becomes relevant (Gursoy and Cai, 2025). By employing predictive analytics, AI could forecast the decay rate of the sensory experiences and suggest optimal times for the re-engagement or reinforcement of sensory cues.

While sensations, feelings and emotions conceptually intertwine, in hospitality, sensations have been examined as memorable sensations (Aksenova et al., 2022) or as emotions. In hospitality, research has identified challenges of nuanced emotions, single measures, self-reporting and conceptual simplicity. Hence, research is encouraged to enhance the view by measuring emotions with technological instruments (Tuerlan et al., 2021). Another view is “neuro-hospitality” and how AI techniques assist in exploring the effects of emotions in hospitality contexts (Doborjeh et al., 2022). Accordingly, sensations in hospitality research can draw knowledge from SM by expanding on both positive and negative sensations, providing insight into why certain services fail or which less apparent external cues cause stress. However, to explore complex emotions, research needs to look beyond sensory attributes and outcomes, such as tourist behaviour (Zhang et al., 2021). For example, real-time sentiment analysis algorithms can interpret customer feedback across platforms, predicting emotional responses to sensory stimuli. A case in point might involve using affective computing (Yi et al., 2025) to modify a restaurant’s lighting and music based on the visitor’s detected mood over time. However, this is largely still unexplored. Hence, we postulate:

The challenge with sensory overload lies in identifying individuals’ sensitivity to stimulation. A persistent discussion concerns creating optimal stimulation in various environments (Spence et al., 2014). The variability in the overload presents methodological and ethical challenges in measuring and creating optimal sensory environments that cater to diverse consumer segments. This complexity provides opportunities for further research by using ML to predict and adjust the levels of sensory stimulation in real time. The issue persists in adhering to a broader target audience regarding niched preferences. A potential solution in physical settings is to schedule sensory intensity in the servicescape, such as lowering lights and music during certain hours to cater to neurodiverse segments (Quinine, 2024). However, more research is recommended to use AI and SM to develop more precise patterns for scheduling hospitality visits.

Sensory overload has been extensively discussed in hospitality, particularly regarding choice and information overload (Sharma et al., 2024). Travel fair studies stress the need to manage atmospheric cues like volume and intensity to avoid overwhelming visitors (Sihvonen and Turunen, 2022). However, more research is needed in hospitality settings. Managing sensory cue intensity is key, as individuals differ in sensitivity, measurable through factors like NFT or optimal stimulation levels (Peck and Childers, 2003). Yet, hospitality often adjusts environments based on general target groups, not individual preferences.

Understanding individual sensory preferences and sensory thresholds, such as volume in lobbies or restaurants, scents, crowding, information and visual stimuli, assists hospitality practitioners in designing more pleasant experiences (Sihvonen and Turunen, 2022). Unlike digital settings that immediately adjust to personal preferences and physical destinations, hospitality environments face barriers to real-time adjustments due to technological and physical limitations. However, as previously discussed, sharing data and implementing AI and ML-based techniques provide new insights into these challenges. Practically, AI-enabled real-time personalisation techniques, such as real-time sentiment analysis or physiological tracking, can gauge the level of sensory stimulation and adjust the intensity of lighting or music to prevent overload (Li et al., 2018). In addition, IoT devices like smart thermostats and lighting systems are used to create personalised guest experiences by integrating guest profiles to enhance sensory experiences (Yi et al., 2025). By leveraging AI to analyse sensory data, hospitality practitioners understand guest preferences and develop strategies to adjust sensory cues dynamically. Addressing these issues would assist previous theoretical and practical challenges regarding sensory overload (Fong et al., 2023) and facilitate sustainable consumer choices (Sharma et al., 2024). Therefore, the following is proposed:

Our study highlights the transformative integration of AI into SM within the hospitality domain. Through an extensive literature review and thematic analysis, we have identified key challenges and opportunities across various facets of SM, including atmospheric and sensory cues, (in)congruence, multi-sensory cues, perception, sensations and sensory overload. Moreover, we underscore the transformative potential of AI in SM by introducing practical, technology-driven solutions that directly address identified challenges. Deep learning models, such as CNNs, can analyse visual aesthetics in hospitality settings to optimise spatial design and décor. Similarly, NLP models can process guest feedback to identify sensory preferences and adjust services dynamically. Reinforcement learning algorithms offer a robust framework for optimising multi-sensory experiences, such as dynamically adjusting lighting or music in real time to match guest moods detected through sentiment analysis. Generative AI, such as generative adversarial networks, can simulate and test new SM strategies in virtual environments, enabling hospitality businesses to pre-evaluate the impact of sensory combinations before real-world implementation. Therefore, the hospitality industry can create adaptive, personalised experiences that align with consumer expectations, enhance satisfaction and drive brand loyalty (Wang and Uysal, 2024).

SM-informed AI enhances hospitality by delivering immersive sensory experiences in both physical and digital spaces. It provides insights into consumer perception, reduces bias and adjusts sensory inputs to avoid overload and improve satisfaction. This paper identifies key themes and proposes future research directions, encouraging scholars to test AI tools and address ethical challenges, including psychological impacts, biometric data use and cultural bias (see Table 1).

Following the recent call for research bridging sensory cues, hospitality and AI (Pandey and Tripathi, 2025), the revealed themes and discussed issues contribute with novel insights into how existing SM frameworks (Krishna, 2012; Spence et al., 2014) benefit from bridging SM with AI in hospitality. By revealing relevant themes, including various conceptual and methodological issues (see Tables 1 and G1), the discussion emphasises that AI provides a deeper understanding of dynamic and sensory experiences, advancing existing SM theoretical frameworks. For instance, the intersection of SM and hospitality research is dominated solely by quantitative and qualitative methods. Employing methodological innovations, such as AI, advances SM research by analysing multimodal data (e.g. multi-sensory cues) that allows for precise modelling and understanding of consumer behaviour and sensory experiences.

The array of empirical evidence derived from such solutions enhances and reconfigures past SM- and S-O-R-influenced models in understanding the role of sensory experiences for hospitality (e.g. the synchronisation of multi-sensory cues), previously unattainable due to methodological limitations. Integrating AI methods propels SM towards a real-time adaptive system, challenging the traditional static view of sensory cues and paving the way for designing a dynamic and personalised sensory experience. Such a shift necessitates a theoretical re-evaluation, recognising sensory cues as flexible and adaptable elements (Eklund and Helmefalk, 2022) in a dynamic consumer environment, rather than fixed entities (Krishna, 2012). By showcasing examples in Table 1 and discussing future avenues within these contexts, we argue for a potential starting point for intermixing these methods and tools.

Existing SM research explains how to effectively influence consumer behaviour at a subconscious level, but has limited scope (Pandey and Tripathi, 2025). AI enables the analysis of complex and large-scale sensory data, identifying subconscious consumer responses to sensory stimuli that otherwise remain undetected in traditional research paradigms (Fong et al., 2023). Such data-driven insights advance SM and hospitality research an understanding of how to influence consumers and vistors at a subconscious level more effectively in hospitality, leading to a real-time, dynamic and personalised sensory experiences.

Traditional SM theories often lack generalisability due to the contextual nature of sensory experiences. AI helps predict and generalise these experiences across various hospitality settings, including cultural, environmental, and service contexts. This cross-contextual approach refines SM frameworks, making them applicable from physical to virtual environments (Manis and Madhavaram, 2023), helping tailor strategies for different settings. As research stresses the need to adapt theories using AI insights (Calderón-Fajardo et al., 2024), this study highlights the challenge of applying past findings to new contexts, such as shifting from retail to hospitality or virtual to physical tourism. Our findings outline pathways for future research on how AI can define SM elements and predict consumer behaviour and responses.

This study shows how AI can enhance SM in hospitality by automating music, lighting, and scent based on real-time customer emotions, boosting satisfaction, loyalty and engagement. To guide managers, we propose an AI integration framework inspired by Kim et al. (2025), using a restaurant visit as an illustrative example. For additional illustrative cases, see Section A.2 of the supplementary file:

1. Audit sensory touchpoints (lighting, music, aromas, food presentation, textures) using guest feedback and behavioural data.

2. Identify AI opportunities, such as adaptive playlists and smart lighting/scent systems.

3. Implement AI tools like music that adjusts to crowd levels or lighting that shifts from energetic at lunch to relaxing at dinner.

4. Train staff to explain these features to guests, enhancing personalisation.

5. Continuously refine SM using data on guest behaviour and sentiment analysis.

Implementing AI in SM within hospitality involves organisational and economic challenges. High upfront costs for technologies like IoT devices and sensors, along with compatibility issues with legacy systems, can be limiting especially for small businesses. A lack of technical expertise or staff resistance also requires investment in training or specialised hires (Ruel et al., 2021). However, assessing internal systems for data access and integration across the guest journey is crucial. As data evolves into a dynamic commodity, sensory alignment may shift over time. Table 1 outlines strategies to help practitioners address these challenges.

AI’s growing role raises ethical concerns, especially around biometric data like facial recognition, emotion analysis and physiological feedback. Strict compliance with data protection laws (e.g. GDPR) is essential to protect consumer rights. Biased AI models can reinforce stereotypes, making diverse data sets, bias audits and stakeholder collaboration critical (Brey and Dainow, 2024). Ethical AI use in SM requires industry standards for data handling, sensory adjustments and transparency (European Commission, 2019). Consequently, managers must adopt frameworks like data protection impact assessments for biometric and emotional data, as well as the fairness, accountability and transparency principles to minimise algorithmic bias. For instance, ML systems that personalise lighting or music based on customer profiles and physiological signals (e.g. heart rate or skin conductance) must include clear opt-in consent, anonymisation protocols and explainability tools to safeguard consumer autonomy. Explainable AI could also ensure transparency of how sensory personalisation decisions are made. In addition, AI-driven sensory simulations using VR/AR should follow inclusive design principles to ensure accessibility and equitable experiences across diverse guest profiles.

With comprehensive updated future data sets, AI assists in understanding sensory preferences across different regions and cultures, paving the way for more ethical (e.g. sensory overload), inclusive and resonant marketing strategies.

The supplementary material for this article can be found online.