How Brand Websites Compete With AI Agents

A shopper getting ready for their first overnight camping trip can now ask an AI agent what to buy and get a thoughtful answer in seconds. The brand may still receive the order while losing the conversation that shaped it: the customer’s priorities, their objections, and the reasoning behind their choice.

Today, a brand’s website must earn its place in the customer journey. That means helping people make better decisions and accomplish more than they could through a conversation elsewhere. A polished catalog and an efficient checkout are only part of that experience.

Consider the choice from the customer’s perspective. On a conventional website, they translate a need into categories, apply filters, open product pages, and work out how the options fit together. In a conversation with an agent, they can begin with what they are trying to accomplish. They can explain their circumstances, ask follow-up questions, and work through uncertainty in their own words.

If a brand’s own site can’t offer comparable guidance, customers have every reason to take their questions somewhere else.

We see three connected changes: an interface that combines navigation, conversation, and real-time interactive components; discovery organized around customer goals; and product knowledge rich enough to support those interactions reliably. Each depends on the others, and a Shopper’s Profile ties them together with an editable record of the customer’s goals, preferences, and circumstances.

Navigation gives people familiar entry points and the freedom to browse. Chat lets them describe a goal, clarify a concern, or challenge a recommendation. An interactive canvas brings together relevant content and components in real time, giving them something to compare, configure, or change as the conversation develops.

These modes share context. A question can reshape the canvas, and a selection in the canvas can inform the conversation. Someone can examine a product, compare alternatives, and return to their plan without explaining their needs again. Navigation is still there for customers who already know what they want or just feel like exploring.

The interface evolves with each decision. A broad question might produce a few possible directions. A specific concern might bring up a comparison. A choice might update a complete kit and its cost.

Adding a chat window is only a starting point. The bigger opportunity is letting the conversation change what the customer can see and do.

Imagine someone visiting Windridge, an outdoor retailer that carries brands like Patagonia and Arc’teryx, with a goal: “I want to prepare for my first overnight camping trip.”

A category-based site asks that person to choose among footwear, jackets, packs, and accessories. That structure works when they know what to buy. A beginner is often still figuring out what matters, which items work together, and what they already own that will do the job.

A goal-led experience can begin with “What’s your next challenge?” The customer describes a cool-weather hike with a chance of rain, says they already have boots but need everything else for an overnight stay, and sets a gear budget of $700. The guide asks the few remaining questions that would change its advice, like how strenuous the walk will be and whether they tend to run hot.

A simple request, “Help me prepare for an overnight hike with a friend,” becomes a guided shopping experience organized around the trip. Navigation, an ongoing conversation, and a curated product canvas work together to help the shopper build a complete kit, starting with the biggest decisions (shelter and a pack) before moving through sleep systems, clothing, and smaller essentials. The guide explains why each item matters, asks about missing details such as the overnight forecast, and encourages the shopper to reuse gear they already own.

They can explore products, check fit, add what they need, and save the plan to their profile to pick up where they left off.

When the shopper opens the Trail Rain Shell, the experience connects it to their overnight plan and the fleece they already own. The guide explains what the shell is for, helps them check fit and layering, and points out what to confirm before choosing a size. If they ask about alternatives, the canvas shifts to a side-by-side comparison of price, coverage, packability, and room for layers, and the guide says which one it would start with and why.

A helpful conversation establishes context, and the next visit should be able to build on it. We call this shared record the Shopper’s Profile: a visual, editable document that captures what the customer wants to achieve, what matters to them, and what they already own. It gives the shopper a clear view of what the experience remembers and uses to guide recommendations.

Anything the guide infers stays a suggestion until the shopper confirms it, and shoppers can edit, pause, or remove what’s saved.

The profile should have a visible effect on the journey. Change the budget and the kit updates. Confirm an owned fleece and the experience builds around it. Ask why a jacket was suggested and the guide can point to the goal and preferences behind it.

This experience requires more than a new interface. A conventional product record can tell the system that a jacket is waterproof, available in four colors, and priced at $180. Reliable guidance also needs to know the uses it supports, its limitations, and how it relates to the rest of the customer’s gear.

Those relationships need to be explicit. A complementary insulating layer, an alternative shell, and an accessory required for a particular use are different kinds of recommendation. Lumping them together as “related products” leaves the system without the reasoning it needs to put together a useful kit.

Each claim needs a source (manufacturer specifications, reviewed expert guidance, or customer feedback), a date, and the models it applies to, so teams can keep that knowledge accurate. Where information is missing, the experience should ask or admit the gap.

Recommendation rules connect the customer’s context to this product knowledge: what must be checked, which constraints take priority, and when the system has enough information to suggest a next step.

The aim is to encode the judgment of a knowledgeable outfitter precisely enough that the experience can apply it consistently. That takes expert review and ongoing ownership. Ingesting existing product copy can start the work, but it won’t supply the missing expertise on its own.

A brand’s own website has the chance to combine this guidance with its product knowledge, interactive tools, live inventory, and service. A customer could explore a complete kit, understand each recommendation, confirm availability, and move into purchase or human help without losing the thread.

External agents can also offer rich interfaces and connected actions. A brand therefore has to show its advantage through the depth and usefulness of its experience. For the outdoor shopper, that might mean a better explanation of layering, a more practical comparison, or continuity between online guidance and an in-store fitting.

Brand expression belongs in those interactions. It shows up in the questions the guide asks, the trade-offs it explains, the products it presents, and the care it takes with the customer’s constraints.

Brands should also make accurate information available to external agents, because some customers will prefer those channels. Being present there can sit alongside a direct experience that gives people good reasons to visit and come back.

A practical starting point is one focused journey, such as preparing for a first overnight trip. Define what a successful outcome looks like, identify the product knowledge it requires, and build the components that help customers make the important decisions. Add a simple profile for the goal, constraints, and existing gear, with clear controls for saving and editing. Connect the journey to current commerce data, and give a team responsibility for its guidance after launch.

IXP is the foundation for this work: structured knowledge, shared customer context, AI composition, and a governed component library. The design system defines how components behave and work together. Governance determines which claims and recommendations are allowed, how uncertainty is handled, and when a person needs to step in.

Then test the business case. Compare the new experience with the existing journey, and measure revenue per visitor alongside margin, returns, and whether customers could finish what they came to do. With the right permissions, the questions people ask will show where knowledge is missing and which choices confuse them, and those findings should shape the next round of improvements.

Customers will decide where the relationship develops. Brands have to build an experience worth choosing.

Fantasy builds IXP experiences with brands, starting with one customer goal like this one. If you’d like to see what yours could look like, let’s talk.