AI-powered Digital Sales Assistants

With the rise of AI, an increasing spread of digital sales assistants can be observed. From the customer's point of view, which features of digital sales assistants are actually most important, and how do these features influence shoppers' willingness to follow product recommendations?

• Among digital sales assistants, consumers prefer those with a humanized and likable appearance rather than artificial ones.
• Preferences for assistant features vary between in-store and online settings regarding appearance. In-store interactions benefit from humanized, credible cues, whereas at-home interactions are more effective with less humanized yet more likable digital sales assistant designs.
• Personalized, engaging conversations are crucial for influencing shopper decisions, while influencer-based information sources are ineffective.

Imagine a digital assistant that not only answers questions but also understands your needs, offers personalized recommendations, and makes your shopping journey smoother and more engaging. AI-powered conversational agents, often seen as chatbots or virtual assistants, are revolutionizing the way brands interact with customers. These smart systems simulate humanlike conversations, using language, gestures, and even expressions to create an experience that feels more personal and engaging than ever before.

Creating successful digital sales assistants goes beyond a one-size-fits-all approach. Just like human interactions, these agents must align with user expectations and adapt to how shoppers interact with digital media across websites, apps, or physical stores. Their design must build trust, enhance value, and respect privacy, all of which influence buying decisions.

Research shows the need to tailor these assistants’ features to different shopping environments and customer preferences. Understanding which features resonate in various contexts, such as the relaxed setting of home shopping versus the busy in-store experience, is crucial. In particular, this study addresses three questions:

• Feature Importance: Which features of digital sales assistants are most important for shoppers?
• Impact on Recommendations: How do these features influence shoppers' willingness to follow product recommendations?
• Channel Differences: Do preferences for the features of digital sales assistants differ between in-store and online shopping?

Drawing from an in-depth review of current research, we’ve identified three key feature dimensions that shape the effectiveness of digital sales assistants: nonverbal cues, informational cues, and conversational cues.

Nonverbal cues focus on the assistant’s appearance and embodiment—whether it looks humanlike or artificial and if its design emphasizes credibility or likability.

Informational cues define how information is sourced and presented to the shopper. The assistant might draw insights from AI-generated summaries of reviews, expert opinions, or influencer recommendations. The delivery style also varies; information can be presented through voice alone, voice with a written transcription, or a combination of voice and visuals, each adding a different layer of engagement.

Conversational cues determine the nature of the interaction, such as whether the dialogue feels personalized or follows a standard script. Privacy settings are also crucial, with options ranging from deleting conversation data after the interaction to using it for further training.

To gain insights into how consumers interact with digital sales assistants, a team of researchers from NIM and Technische Hochschule Ingolstadt conducted a study involving 446 participants recruited from Prolific, a widely used online research platform. Participants were randomly assigned to two shopping environments: in-store shopping or a home online shopping experience. We used visuals to bring these shopping environments to life, illustrating the look and feel of each digital sales assistant interaction. To illustrate how digital sales assistants present information, we used real summaries of reviews, expert advice, and influencer commentary, all anonymized to prevent bias.

Using an advanced statistical design, participants compared different digital configurations of sales assistants and were shown their most and least preferred options. We then measured how these digital sales assistants influenced their purchase intentions. This approach allowed us to identify which features drive the most engagement, offering valuable insights for marketers looking to optimize digital sales assistants in various shopping environments.

Our study reveals that consumers prefer digital sales assistants with a humanized and likable appearance rather than artificial, purely credible ones. When it comes to the source of information, advice from human experts is the clear favorite, outperforming AI-generated review summaries, while influencer advice is the least preferred. The most engaging presentation combines voice with visuals, making the information feel more dynamic and accessible. The style of conversation also plays a significant role: Personalized digital assistants that delete conversations after use are favored over those that retain data for training purposes, reflecting growing consumer concerns about privacy.

Interestingly, preferences for assistant features vary between in-store and online settings regarding appearance. In-store interactions benefit from humanized, credible cues, whereas at-home interactions are more effective with less humanized yet more likable digital sales assistant designs.

When it comes to what matters most, the source of information is the top priority, influencing 36% of consumer decisions. Privacy features are also highly valued at 20%, highlighting the importance of data security, while the assistant’s social appearance is the least significant factor at just 2%. Our advanced modeling shows individual preferences vary widely, especially in nonverbal cues and the emphasis on expert advice, underscoring the need for tailored digital assistant designs that meet diverse consumer expectations.

Our study not only identified preferences for digital sales assistant features but also explored how different configurations impact purchase intentions. Paticipants rated their likelihood of following recommendations from assistants designed to be either highly appealing or not appealing at all. The analysis revealed that only a small portion of the variation in purchase intentions could be explained by the features examined, highlighting the complexity of consumer behavior. Key findings suggest that personalized, engaging conversations are crucial for influencing shopper decisions, while influencer-based information sources are ineffective.

• Marketers: To maximize the effectiveness of digital sales assistants, retailers need to equip them with reliable information sources and ensure clear communication about data sourcing and privacy practices. Personalized expert advice, complemented by visual tools like product comparison tables, enhances shopper utility. However, subjective influencer recommendations should be avoided in the final purchasing stages, where a focus on practical, goal-oriented support is more effective. Preferences for nonverbal cues vary between at-home and in-store interactions. Retailers can cater to this variance by offering customizable nonverbal features in at-home settings, allowing shoppers to choose based on personal preferences and the flexibility of time. In contrast, in-store digital assistants should maintain a humanized and credible appearance, reinforcing expert-level advice and creating a consistent, trustworthy shopping experience.
• Consumers: Consumers stand to benefit from more tailored and effective interactions with digital sales assistants. With digital sales assistants offering personalized and expert advice, coupled with clear and honest information, shoppers can make more informed and satisfying purchasing decisions.
• Society: At a societal level, the evolution of digital sales assistants can lead to more informed and empowered consumers. As technology becomes better at meeting individual needs and respecting privacy, it contributes to a more positive and efficient shopping environment. This shift can foster a more ethical and consumer-centric marketplace.