AI in retail (part 1): Better experiences, but bigger risks

Artificial Intelligence (AI) has entered our daily lives incredibly fast, including in the way we browse, choose, and buy products. The integration of AI-powered personalization, intelligent search, customer support automation, and predictive analytics into the online shopper’s journey is allowing companies to innovate, optimize processes, and enhance customer interactions. However, this increased power brings complex challenges. As AI adoption increases, so does the awareness of its limitations. Their outputs, while powerful, can sometimes lack the context and empathy that define a truly human-centered experience, and this is where a solid, value-based strategy is essential.

There are 3 key areas where AI drives impact: customer experience, operational efficiency, and employee productivity. In Part 1 of this article series, we will explore how AI can elevate the e-commerce customer experience and the risks it can create when not properly implemented.

What is the current state of AI in e‑commerce?

Companies are investing and seeing results

A recent study shows that 91% of retailers are now engaged with AI. Having 58% actively deploying AI, while 33% are assessing AI solutions. Most companies have reported that AI is reducing annual costs and positively impacting revenue, and plan to increase investment in 2026. AI Agents are currently at the top of investment lists, used to accelerate operations, enhance customer experiences, and improve decision-making.

Consumers are adopting AI with caution

Shoppers are already using ChatGPT, Gemini, and Perplexity not just to figure out what to buy, but to make purchases through new checkout experiences, but only 34% are willing to let AI assistants make the purchase on their behalf. When it comes to customer assistance, most customer questions are now being resolved without human intervention. Moreover, 72% of consumers are open to AI-powered interactions, either when available or with the option to transfer to a live agent if needed.

Big opportunities for customer experience

AI-powered personalization that feels personal

We are moving from basic segmentation to real-time 1-to-1 personalization. By combining machine learning, natural language processing, and behavioral data, it’s now possible to predict individual shopping needs and preferences. This means it will not only be easier for the customer to find what they are looking for, but also to discover what they didn’t even know they needed. Products can also display unique descriptions, images, and marketing tailored to the specific tastes of the customer.

A great example of a brand that uses AI to offer better experiences is Sephora. They use AI to provide highly personalized product suggestions, taking into account factors such as skin type, past purchases, and customers’ beauty goals. They also provide several AI tools to enhance the online shopping experience. Customers can upload a selfie to get recommendations for the perfect makeup shade or use virtual try-ons to see how products look on their face before buying, reducing the uncertainty of online shopping. They have also recently announced an App in ChatGPT for customers to get AI-powered guidance.

The rise of AI customer support

Today’s customers expect real-time support and are less tolerant of delays. Customers usually prefer to talk to a person, especially when it’s an urgent matter. Unfortunately, human assistants are not always available, and basic chatbots are usually not that helpful. AI assistants can provide a middle ground because they are more intelligent than a typical chatbot and can support customers when no humans are available. Companies can now provide 24/7 support in multiple languages and manage their resources in a more efficient way.

IKEA launched an AI chatbot that could handle almost 50% of customer requests. Instead of simply cutting costs, they focused on providing better service to customers and trained their available employees to handle more complex topics, such as interior design advice, resulting in $1.4 billion in additional revenue. 

Even better than having assistance always available is the opportunity to provide proactive service. AI can help detect, for example, whether a delivery is likely to be late and act before the customer even notices. This provides brands with opportunities to show they care about their customers. 

Search boosters

Visual search, semantic search, and AI assistants are all ways to allow customers to find the right products faster. Due to conversational and context understanding, customers can search for things they need without knowing exactly what or how to describe them.

By using semantic search or a shopping agent, customers can search for items like “black tie dress for wedding guests” or “best laptops for graphic designers” without having to go through a long list of filters or read dozens of reviews. Instacart is a North American online grocery delivery and pickup service that released an AI-powered search that allows customers to ask questions such as “What’s the best fish for tacos?” or “What can I use in a stir fry?”, supporting users in making decisions about budget, dietary specifications, cooking skills, personal preferences, and more.

Visual search is also a great way to find similar products without having to describe them or know the exact brand name. Secondhand lovers know how time-consuming it is to find the perfect items. Vinted, the secondhand marketplace, now allows users to upload a photo of an article they like and search for similar items. This feature allows customers to find what they are looking for much faster.

Omnichannel

Nearly 60% of customers say their preferred service channel depends on the situation, which means that shoppers expect brands to be available on multiple channels. And whether in a physical store, on WhatsApp, Instagram, or on an e-commerce website, customers expect a seamless experience. AI ensures the interaction feels like a single, continuous conversation by aggregating and analyzing unstructured data across every touchpoint to maintain real-time context.

Brands like Sephora use AI to bridge digital and physical worlds. A customer can find a makeup shade in the app, and when they go to the store, the advisor can immediately access their data to provide a tailored, customer-centered consultation, so the customer doesn’t need to repeat information.

While omnichannel isn’t a new concept, AI allows brands to deliver even more seamless interactions across all customer touchpoints. Traditional omnichannel is reactive: it shows what the customer did. With AI, it can aggregate data from every channel and monitor customer sentiment in real time to predict what customers need next.

Accessibility

AI lowers the cost of making experiences accessible. With real-time translation, voice interfaces, adaptive UIs, and simplified language, brands can now reach a wider audience. Augmented and virtual reality can enhance sensory experiences and provide real-time information. It’s a major opportunity for companies to integrate assistive technologies to enhance interaction and communication.

AI-powered assistants, for example, can assist users with impairments by answering product queries, helping them find information, or even performing advanced actions such as creating orders without requiring the user to manually complete the steps.

Another opportunity for e-commerce is to use real-time UI personalization, which happens when the system analyzes preferences, devices, and interaction patterns to adapt the interface to the user’s needs. For example, by detecting that a user is struggling, it can automatically provide a simplified checkout flow, bigger fonts, larger touch targets, or voice search. This could bring major improvements for users with limited vision, cognitive, or motor challenges.

The risks

Too much personalization

Personalization can be very useful, but it can feel “uncanny” if it starts performing too well. Many people mention having been in a situation where it feels like their smartphone is listening to them. As far as we know, they are not. Companies have so much data about us that it feels like they know us. This can lead to a feeling of being under surveillance and leave customers uncomfortable. Too much personalization can also create “filter bubbles” where users are not exposed to new content, which can lead to social alienation. 

New 2026 regulations (like the updated EU AI Act) now require AI-generated personalized content to be labeled. Failing to disclose that a video or message was “synthetically tailored” for a specific user can result in massive fines and permanent brand damage.

Lack of control

It’s known that AI can confidently produce wrong information. In a customer context, this breaks trust and can create real liability. When we delegate decision-making to AI agents, we risk generating misaligned or inaccurate content at scale. That can undo years of brand equity in a single interaction. If an agent makes false claims, surfaces outdated controversies, or uses a tone that doesn’t match our “warm-professional” identity, it creates a “trust gap”. Without human oversight and strict governance, these unpredictable outputs can make a brand appear cold, insincere, or even deceptive, ultimately affecting customer trust.

Loss of human touch

Conversations with AI can fall into a cold, overly corporate tone that feels distant or rigid. If an AI tool lacks a human-centered design, it can become a barrier. Not every interaction should be fast and automated. Complaint escalation and complex decisions require human judgment and empathy. Companies that chase efficiency by automating everything can cause bad experiences at the moments that matter the most. Studies show that 40% of shoppers are frustrated by the lack of human support in customer service, and 86% said that empathy and human connection matter more than quick responses.

When efficiency looks cheap

We’re seeing brands increasingly use AI to cut advertising costs, from billboards to full-scale commercials. While for some people this might come unnoticed, others see it as a lack of effort and as a “tacky” way for companies to save money. Especially for big brands like Coca-Cola, which faced backlash over their AI-generated Christmas ad. Consumers expect big brands to invest in authentic storytelling rather than automated outputs.

On the other hand, some companies are adopting a “No AI” movement to emphasize human-made craftsmanship and differentiate themselves, much as a luxury car or a high-end device is marketed on its engineering excellence. If we prioritize cost-cutting over delivering real value and outcomes, we risk looking impersonal and detached, which ultimately alienates the people we’re trying to reach.

Things to watch out for

What AI gets right

• True personalized product discovery, tailored offers, “you’ll also like” that actually works.
• Fast, always-available support that resolves simple issues without friction.
• Smarter search that understands intent, not just keywords.
• Collecting and analyzing data for a seamless omnichannel experience.

Where it goes wrong

• Chatbots that feel robotic, damage brand voice, and frustrate customers when dealing with complex issues.
• Personalization that crosses into feeling intrusive or manipulative.
• AI is deployed to cut costs rather than serve customers.
• Opaque recommendation engines that create filter bubbles, hiding options the shopper would have wanted.

The deeper risks

• Eroding trust when AI gets it visibly wrong (a bad recommendation, or a tone-deaf response).
• Over-automation by removing the human touchpoints that build loyalty.
• Data privacy concerns, especially as AI systems require ever more personal data to function well.
• The expertise gap: companies are adopting AI faster than they can implement it thoughtfully.

Balancing automation with empathy

The goal should be to ensure AI supports the user without replacing the “human” element of the brand. We need to be intentional about where we automate and where we provide a direct line to a person. AI should be a partner in the user experience, not just a tool for optimization.

There is a sweet spot when it comes to AI. Just because it has risks doesn’t mean companies shouldn’t invest in it. What’s important is the motive behind each new feature, and not forgetting that the customer is the most important thing for each business.

A few key points to remember:

• Design human-AI interactions intentionally.
• Transparency about what AI is doing and why.
• AI that amplifies brand voice rather than replacing it.
• Measure success by customer satisfaction, not just operational efficiency.

Companies prioritizing strategic investment in AI will aim to differentiate themselves by creating genuinely good, personalized, and proactive customer experiences, viewing AI as more than a tool for efficiency. On the other hand, businesses with an AI strategy focused solely on reducing costs risk alienating customers by implementing automation primarily for labor substitution, without prioritizing the customer journey.

Ultimately, the market will distinguish and reward companies committed to high-quality, ethical AI experiences, while exposing those whose use of AI is limited to short-sighted expense-cutting. This period will clearly demonstrate a company’s commitment to its customers and its long-term vision.

What comes next?

Are you ready to modernize your customer journey with a strategy that puts people first? Let’s collaborate to build something meaningful.