The Trend Report: The New Rules of AI-Powered Personalization

Why relevance and trust now define personalization

Personalization used to be easy to spot.

It was your name in an email subject line. A row of products labeled “You might also like.” A discount after you abandoned your cart.

Sometimes it helped. Other times, the internet learned one thing about you and made it your entire personality. Buy one blender, and suddenly every website assumes you are opening a smoothie shop.

AI is changing both the sophistication and the stakes of personalization.

Rather than relying only on broad segments or purchase history, brands can interpret what shoppers are searching for now, which products they are comparing and what they are trying to accomplish in the moment.

That creates an opportunity to make ecommerce more relevant and intuitive. It also creates more ways to get personalization wrong.

Research from BigCommerce and Future Commerce found that more than three-quarters of respondents across every age group believed irrelevant product recommendations were worse than receiving no recommendations at all.

Customers do not simply expect brands to recognize them. They expect brands to help them make decisions with less effort.

From customer profiles to customer intent

Traditional personalization starts with what a brand already knows.

A shopper purchased running shoes, browsed outdoor gear, and opened several fitness emails. The brand places them in a segment and serves them more of the same.

But people do not stay neatly inside predefined categories. The runner might be shopping for a baby shower gift. A value-focused customer may suddenly care more about delivery speed because they need something tomorrow.

AI allows brands to respond to immediate signals rather than relying exclusively on a fixed profile. Browsing patterns, comparisons, and conversational prompts can reveal what a shopper needs right now.

In a Forbes Technology Council article, Vinod Sivagnanam points to clickstream data, including browsing patterns, page interactions and the time shoppers spend considering products, as a valuable source of context for AI-powered experiences. He argues that personalization has quickly become table stakes in ecommerce. The real challenge for brands, then, is no longer whether to personalize. It is whether that personalization is perceptive enough to be useful.

The opportunity is not simply to produce more personalized content. It is to create an experience that adjusts as intent becomes clearer.

“The best personalization doesn’t feel like personalization. It just feels like the brand gets it. AI can help brands respond faster and with more context, but there’s a fine line between being helpful and being creepy. The winners will be the ones that use AI to remove friction, not show off how much data they have.”

— Al Williams, Vice President of Market Strategy, Commerce

The search bar is becoming a conversation

For years, ecommerce search asked shoppers to translate what they wanted into a few keywords. A customer might type “blue dress,” scan hundreds of results, and slowly add filters for size, price, and availability.

AI-powered search can begin with the outcome instead:

I need a dress for an outdoor wedding in Nashville. It needs to work in humid weather, fit in a carry-on, and cost less than $200.

That prompt contains context, preferences, and constraints. A conversational interface can interpret them together and narrow the choices faster. This shift is already showing up in real shopping experiences. Amazon’s shopping assistant can tailor answers using a customer’s activity and the context of their request, while Sephora uses AI analysis and virtual try-on data to personalize recommendations and connect shoppers with additional services. Semantic search is also helping retailers move beyond exact keyword matches, allowing a request like “running shoes for wide feet, trail use, under $120” to return relevant options even when those precise words do not appear in the product title.

In the BigCommerce and Future Commerce study, 58% of respondents said they liked seeing AI-driven search integrated into standard search. Nearly half wanted more AI-powered search that could help them choose products using multiple criteria.

This reflects a broader shift identified in the report: brands must become “omnimodal.”

Being omnichannel means connecting experiences across places. Being omnimodal means supporting the different ways people now interact with commerce, including text, images, voice and AI assistants.

The storefront will not always be the beginning of the journey. Personalization must work wherever a decision is being shaped.

Better personalization starts with better data

Conversational shopping can feel effortless, but AI needs enough information to understand both the request and the available products.

To recommend the right dress, it may need details about material, fit, care, inventory, and delivery timing. A shopper looking for a “pet-friendly sofa for a small apartment” needs information about dimensions, fabric durability and stain resistance.

AI cannot personalize around information it does not have.

That makes product data a critical foundation. Titles, descriptions, images, specifications, and attributes must be accurate and detailed enough for both humans and machines to understand.

Customer data matters too, but more is not always better. A preference provided directly by a shopper may be more useful than an assumption drawn from months of browsing behavior.

“A shopper who types, “What suit should I wear to a summer wedding in New York in August?” into an AI engine isn’t searching for a product. They’re asking for a recommendation. Brands whose product data can answer that question, whose catalogs contain the contextual, conversational attributes that map to how humans actually ask, will show up.”

— Michael Scholz, Vice President of Product, Commerce

The goal is to connect the right customer context with the right product information at the right moment.

Customers expect a fair exchange

Personalization requires information, but customers expect something useful in return.

The BigCommerce and Future Commerce report describes this attitude as “data for you, deals for me.” Seventy-four percent of respondents had opted into and out of brand messages on the same day to receive a discount.

That does not mean every personalized interaction needs a coupon. It means the benefit should be clear.

In the study, 58% wanted more abandoned-cart emails when they included a significant discount. Only 16% wanted more emails that simply reminded them to complete a purchase. Respondents also wanted brands to connect online and in-store purchase histories and provide recommendations based on past purchases.

The difference is usefulness.

A reminder that repeats what the customer already knows creates noise. A relevant discount, replenishment suggestion or compatibility warning can make personalization feel worthwhile.

The same principle applies to data collection. Sixty-three percent of respondents had abandoned a cart when guest checkout was unavailable, while 58% had walked away when both an email address and phone number were required for a promotion.

Customers are not necessarily rejecting personalization. They are rejecting a bad bargain.

In that same Forbes article, Sivagnanam also cautioned brands to balance personalization with privacy, be transparent about how customer information is used, and avoid overwhelming shoppers with recommendations. That last point matters the most. AI may give brands the ability to personalize nearly every interaction, but that does not mean every interaction needs it. Sometimes the most customer-centric choice is to make the experience simpler and leave a little room to browse. 

One research participant summarized the standard brands should aim for:

“I don’t care how ‘personalized’ something is, I’m interested in a smooth experience.”

The best personalization removes work. It narrows an overwhelming selection, remembers useful preferences, and surfaces the right information without adding more pop-ups, questions, or distractions.

“Personalization isn’t about more options; it’s about creating a streamlined experience. The best personalization removes work. It narrows an overwhelming selection, remembers useful preferences, and surfaces the right information. No more pop-ups, questions, or distractions — just clarity. Brands focusing on this will win customer loyalty, as consumers crave ease in their decision-making journey.”

— AL Williams, Vice President of Market Strategy, Commerce

What brands should do now

Start with friction, not technology. Identify where customers struggle or leave, then use personalization to solve that specific problem.

Build around intent as well as identity. What a shopper is comparing today may matter more than what they purchased six months ago.

Strengthen product data. AI needs accurate, structured attributes to understand why a product fits a particular need.

Make the value exchange clear. Ask only for information that improves the experience, and show shoppers the benefit.

Measure effort, not just engagement. More clicks and longer sessions do not always indicate success. Consider how quickly customers find relevant products and make confident decisions.

The final word

AI is making one-to-one personalization possible at a scale brands could not previously achieve.

But one-to-one does not need to mean one algorithm watching everything a customer does. However, it can mean recognizing context, respecting preferences, and making each interaction more useful.

The brands that succeed will not necessarily collect the most information or generate the most recommendations. They will exercise the best judgment. Like a great salesperson, effective personalization notices enough to help and knows when to step back.

The future of personalization is not an internet that knows everything about you. It is an experience that knows when to help and when to get out of the way.