AI is shaping how products are evaluated, and increasingly, how they are purchased, writes Frances Dennis, chief marketing officer at Brandwidth.

Agentic search is accelerating, commerce protocols are evolving and advertising inside conversational interfaces is currently being tested. Even in its relatively infant stage, a significant proportion of customers already prefer generative AI to traditional search.

This will have commercial consequences. Traffic from agentic environments is converting quickly as discovery becomes more conversational, synthesised, and closer to transaction.

There is a real upside to this for customer experience. AI can collapse research time and help customers navigate complicated choices with greater ease. Retailers should be leaning into structured data, clear specifications and authoritative third-party signals that increase visibility inside these environments and achieve efficiency.

But efficiency is only one part of the equation.

The digital experience has already seen what happens when optimisation becomes the strategy rather than the tool. In the early years of SEO, some brands successfully integrated ranking logic into experiences that still felt intuitive while others contorted journeys to satisfy algorithms, gradually degrading usability in the process. Pages ranked, but the experience became heavier and less coherent.

The lesson isn’t to resist change, but to always keep in mind what we’re optimising for.

Today, Generative Experience Optimisation (GEO) is emerging as the next frontier. Retailers are rightly focused on making their content interpretable by large language models (and if they aren’t, they should be). Product data is being structured carefully while FAQs, delivery policies and specifications are being aligned with PR and GEO strategy to ensure trusted non-branded platforms reinforce onsite messaging to boost content relevance.

These changes are fast becoming a matter of digital hygiene. Because if you’re not visible, you won’t be considered.

But being surfaced isn’t the same as being chosen. And it’s certainly not the same as building long-term preference.

Speed isn’t always the end game

Agentic tools summarise and recommend.  In many cases, they still link out to owned websites or apps for transactions. As LLMs experiences improve, this handoff will matter more and more. If an AI assistant highlights a particular feature, price point or sustainability credential, the landing experience must reflect that

immediately and coherently. The customer has to feel continuity, which means that AI strategy and owned strategy can’t sit in separate silos; they need to be designed together, from the beginning.

Simultaneously, we need to acknowledge a near-term reality; in some categories, purchase will be possible inside AI interfaces themselves. This compresses the journey even further and reduces the amount of experience a brand directly controls.

But this level of efficiency isn’t always a good thing.

In many retail sub-sectors, carefully managed friction has always played an important, strategic role. From guided configuration in automotive, shade-matching tools in beauty, fit quizzes in fashion or appointment booking in premium retail, these moments are carefully considered ‘Ikea-effect’ mechanisms that support conversion. They also help to protect margin, reduce returns and reinforce brand positioning. In essence, they shape how a customer feels about a purchase before they’ve laid a hand on it.

If AI-native checkout removes those stages entirely in the name of efficiency, brands face a decision. Do they optimise purely for transactional volume, or do they design their AI participation in a way that preserves context and confidence?

That may mean routing higher-consideration purchases back into owned environments, structuring integrations so consultation and configuration remain intact or ensuring that loyalty benefits, bundled services or post-purchase support are strongest within accounts and apps that brands control directly. None of this is about resisting AI but about setting thoughtful boundaries around what’s delegated, and considering upfront the overall brand discovery experience you want to engineer.

Discovery may fragment, comparison may be synthesised and checkout may happen within a conversational interface. But memory, habit and preference are built over time, often within owned ecosystems.

Websites and native apps remain where most transactions take place. They offer more than just being technical endpoints, instead being environments where retailers shape tone, pacing and reassurance. They are where loyalty programmes live, service is resolved and cross-channel continuity can be designed intentionally. As AI mediates more of the top of the funnel, the strategic importance of these environments increases.

Rethink the purpose of your mobile app

Mobile apps remain underleveraged in several retail sectors (particularly FMCG). Adoption is often weak not because consumers resist them, but because the value exchange is thin.

When apps simply replicate web functionality, they add little. When they deliver genuine utility, integrated loyalty, seamless reordering, location-aware services or intelligent personalisation, they become habitual.

The retailers who navigate this shift well will understand that three forces are now operating at once. They will optimise for algorithmic visibility and design experiences that genuinely empower customers to make confident decisions. Finally, they will protect the journey elements that build long-term brand equity, even if that means resisting frictionless speed in certain contexts.

AI will continue to evolve, interfaces will change and commercial models will mature. But customers will remain human. They’ll still respond to journeys that feel clear, considered and respectful of their time.

The challenge is designing retail ecosystems where both efficiency and experience can coexist, without eroding the brand value that makes being discovered worthwhile in the first place.

Click here to sign up to Retail Gazette‘s free daily email newsletter