AI Can’t Quite Crack the Code of Shopping

Autonomous agents are great at structured tasks. But shopping requires intuition, adaptability, and emotional understanding — things today’s AI still can’t fake.” – Dr. Ilya Polikarpov, Founder of Raga Labs, 2025.

While AI has undeniably made strides in transforming industries like finance and healthcare, it hasn’t quite mastered the art of understanding human desires in the chaotic, often unpredictable environment of shopping. The task of purchasing isn’t simply about finding the best price or the most efficient option, it’s about connecting with personal tastes, emotions, and the context of the moment.

Autonomous agents, designed to handle structured, repetitive tasks, excel in environments with clearly defined parameters. But retail, with its unpredictable customer behaviors, emotional nuances, and vast array of choices, remains an arena where AI has yet to fully shine.

As Michael Parekh pointed out in his recent article, “AI agents are not designed for our messy, nuanced world of shopping.”

Although contemporary AI excels at processing substantial data and making rapid judgments, the dynamic and often chaotic nature of e-commerce presents a challenge. These systems struggle with continuously evolving product suggestions, customer preferences, and real-time modifications.

A perfect example is how autonomous agents manage supply chains or handle data entry. These systems thrive in environments where inputs, processes, and outputs can be clearly defined. When it comes to shopping, however, the landscape is entirely different. Imagine the complexity of purchasing a pair of shoes. Beyond simply comparing prices and features, human shoppers consider various emotional and contextual factors: how they feel that day, what’s trending, how a product fits into their lifestyle, or even the aesthetics of a website. These elements require intuition and adaptability qualities that AI struggles to emulate.

AI agents as “assistants” in retail assumes the customer needs a simple straightforward answer but this approach underestimates the complexity of human shopping habits. When buying clothes or gadgets, customers often seek recommendations that go beyond technical specifications. They want a shopping experience that anticipates their desires and intuitively aligns with their emotions. 

Therefore personalization should go beyond basic algorithms. AI today can recommend products based on past purchases, but true personalization involves understanding a shopper’s deeper motivations. AI agents will need to tap into not just what a person has bought, but why they’re shopping in the first place. Are they looking for something to express their personality? Do they need an item for a special occasion? 

Some startups are incorporating advanced technologies like emotional intelligence, context-awareness, and adaptive learning to make shopping assistants more personalized, intuitive, and responsive to human behaviors.

Perplexity AI‘s shopping agent is turning heads with its innovative approach to e-commerce. By seamlessly integrating product research, price comparison, and direct purchases within its platform, the AI assistant offers a streamlined and intuitive shopping experience. With features like visual search and “Buy with Pro” for direct purchases, it’s pushing the boundaries of how AI can enhance online shopping.
Despite advancements in structured tasks, AI systems frequently find it challenging to understand and react suitably to human emotions. 

Stitch Fix, renowned for its clothing subscription service, uses AI to provide personalized styling recommendations. Its AI system goes beyond predicting sizes and color preferences by analyzing shopping history and stylist feedback, ensuring that the suggestions remain relevant and aligned with a customer’s changing needs. It has entirely redefined the way AI interacts with personal style, blending cutting-edge technology with human intuition to create a shopping experience that feels both personal and effortlessly chic. In a world where many AI shopping agents churn out recommendations based on basic algorithms, Stitch Fix has elevated the art of personalization. Instead of simply analyzing purchase history, their machine learning system digs deeper, learning about each customer’s unique preferences, body type, and style choices. But it doesn’t stop there, it integrates the human touch of professional stylists who refine and enhance those recommendations, ensuring that each piece aligns not just with what the AI thinks of a person like, but with the deeper nuances of the user’s personal taste.

What sets Stitch Fix apart is its ability to create an adaptive learning system. With each interaction, the AI becomes smarter, understanding which styles fit best, what fabrics feel most comfortable, and what types of outfits make you feel your best. It’s a dynamic process, where the more feedback you give, the better the experience becomes. But there’s a critical element Stitch Fix understands that many AI agents don’t: context. Shopping isn’t just about the items one has bought before; it’s about understanding the moments in their life that shape their choices. Whether it’s finding the perfect dress for a wedding or selecting the right business casual attire for a promotion, it tailors its recommendations by tapping into the emotional context of these moments. In a world where AI often seems like it can’t quite capture the subtleties of human emotion and taste, Stitch Fix has cracked the code. It’s a seamless marriage of artificial intelligence and human expertise. 

But having an AI agent for shopping is just not enough as it comes with its own burden, as  Greg Zakowicz, Sr. Ecommerce Expert at Omnisend, noted, “Consumers are open to AI enhancing their shopping experience, but there’s a big difference between receiving personalized recommendations and handing over full purchasing control. Large-scale adoption of AI-driven purchasing requires a fundamental shift in consumer behavior — and that’s not happening anytime soon.”

Data privacy and security are significant challenges when using AI shopping agents. These agents rely heavily on user data to provide personalized recommendations, optimize shopping experiences, and streamline transactions. However, this dependence on data presents several risks and concerns for both consumers and businesses. While this data helps to improve user experiences, it also raises privacy concerns, especially regarding how much personal information is being collected, whether it is stored securely, and who has access to it.

Despite the progress made, AI has not yet cracked the code for fully understanding the emotional and contextual layers that make up the human shopping experience. For now, the ideal shopping experience still involves a mix of human intuition and AI efficiency. AI can streamline the process, provide helpful recommendations, and improve convenience. But when it comes to making those deeply personal, emotionally charged decisions like buying a gift for a loved one, or choosing the perfect outfit for a special event there’s no substitute for human understanding.

We may see AI systems that are better equipped to handle the emotional nuances of shopping, and these systems will likely blend the best of both worlds: the precision of AI and the empathy of humans. Until then, the AI-powered shopping experience remains an evolving work in progress. It’s clear that while AI has great potential to transform the way we shop, there are still many layers of human complexity that it needs to understand—and for now, that’s something only humans can truly deliver.