Experts suggest that the success of AI shopping assistants will hinge on data readiness and security rather than merely speed. Consumer adoption surged during the recent holiday season, with AI-driven tools and assistants influencing 1 in 5 purchases, translating to approximately $262 billion in revenue, according to Salesforce.
Brands that rolled out their own shopping agents ahead of the holiday saw significantly greater growth compared to those that did not. Caila Schwartz, director of consumer strategy and insights at Salesforce, highlighted that the momentum is increasing pressure on customer experience (CX) leaders to act. However, experts warn that the most significant risk lies in deploying these technologies before ensuring organizational readiness.
AI shopping assistants are designed to enhance the shopping experience through personalized, conversational, and efficient customer interactions. Nevertheless, brands must first establish a solid data strategy. Unified data is deemed the most crucial prerequisite; AI agents require comprehensive context about both customers and product catalogs, which are often fragmented across various systems. For instance, product information might reside in a product information management system while inventory tracking exists in an enterprise resource planning system.
“Your outcomes are really as good as the inputs that you provide,” stated Grant Deken, head of product at Klaviyo, an AI marketing and CRM provider. He likened the process of training an AI agent to onboarding a new employee. Brands must also focus on data hygiene, content quality, and proper integrations to achieve meaningful results. Without a united approach to customer, product, and brand data, the personalized experience promised by AI risks collapsing under the weight of inaccurate information.
Instacart’s CTO, Anirban Kundu, cautioned that “if the assistant recommends out-of-stock items, misstates policies or misrepresents compatibility, trust erodes quickly.” He also noted the “cold start challenge,” where limited initial data can hinder performance at a time when first impressions are crucial. While in-house development of AI shopping assistants is feasible, many brands lack the resources for effective implementation. Some companies that initially attempted internal builds found their efforts lacking, describing them as mere “science experiments,” according to Zoovu, an AI-powered product discovery platform.
Security concerns are another compelling reason for brands to partner with specialized vendors. Shopping assistants are susceptible to prompt-injection attacks, where users attempt to manipulate the system to disclose sensitive information or offer free products. “You can’t protect yourself from prompt injection with clever prompts,” Taylor, a representative from the industry, remarked.
Integrating AI into the Customer Lifecycle
To ensure the successful deployment of AI shopping assistants, CX leaders must rigorously evaluate vendors. Beyond demonstrating the capability to drive measurable sales outcomes, vendors should illustrate how their technology integrates into the entire customer lifecycle, rather than treating the AI agent as a standalone tool. Insights gained from shopping interactions, including those from customers who do not make purchases, can feed back into marketing strategies and future engagement.
Deken advised brands to initiate their AI journey with a single product category. “Start in a very scoped, clearly measurable project, and then build a business case from that,” he said. Experts also cautioned against a “set-it-and-forget-it” mentality, emphasizing the need for brands to continually update their knowledge bases and enhance capabilities over time. “Building a reliable assistant isn’t a feature launch, but an ongoing engineering commitment across machine learning, infrastructure, and product,” Kundu explained. He added that maintenance is a continuous endeavor as catalogs evolve, promotions change, and foundational models are updated in ways that affect consumer behavior.
Brands should also explore innovative methods to engage consumers consistently. For instance, Zoovu reportedly tripled engagement rates by prompting consumers with frequently asked questions. Taylor illustrated this by mentioning an electronics retailer that might suggest customers inquire about the number of USB ports on a laptop. “We find that a lot of customers don’t even know what they’re supposed to ask for,” he stated. Providing guided options can significantly enhance user engagement with the assistant.
As adoption of AI shopping assistants remains in its infancy, CX leaders are advised to avoid overwhelming users with technology. Forcing customers who favor traditional search methods to adopt AI tools risks alienating them. The assistant should be accessible without overshadowing the browsing experience that most shoppers still prefer.
Looking to the future, Schwartz predicts that 2026 will mark “the year of the customer-led agentic experience” in retail, highlighting a forthcoming wave of innovation. “We’re going to see a lot of innovation happening in this space,” she said, anticipating developments in agentic shopping inspiration and more cohesive catalogs across various channels.
See also
Bank of America Warns of Wage Concerns Amid AI Spending Surge
OpenAI Restructures Amid Record Losses, Eyes 2030 Vision
Global Spending on AI Data Centers Surpasses Oil Investments in 2025
Rigetti CEO Signals Caution with $11 Million Stock Sale Amid Quantum Surge
Investors Must Adapt to New Multipolar World Dynamics
