Comparison, spec-matching, review-sifting and deal-checking all happen in an assistant. By the time someone reaches your PDP, they've been told what to buy. Your product page is a confirmation screen, not a persuasion surface.
"Best waterproof jacket under £150" now returns three brands and clear reasoning reason. Visibility is a function of how machine-legible your catalogue, pricing and reputation are, and it’s getting harder to buy your way into consideration.
Drip pricing, was/now, confusing pack sizes, bundles nobody fully uses, subscription inertia. Price architecture designed for tired humans doesn't survive an assistant that does the maths and shows its working.
The weekly shop, the reorder, "find me a cheaper equivalent" is all delegated to software optimising on price, availability, delivery slot and past dissatisfaction. Impulse, merchandising and end-of-aisle psychology don't work on an agent.
The buffet becomes a picnic: agents force à la carte pricing into the open and quietly retire the things you don't use. The casual customers who subsidised your bundle economics stop buying it. Bundles survive only where the whole is genuinely cheaper or better than the parts.
Whoever the customer trusts to transact becomes the shelf. Retail's gatekeeper stops being search and starts being the assistant, and in doing so being unreadable to it is the new out-of-stock.
Customers will shop through assistants and expect to transact without leaving them. Expose catalogue, live price, stock, delivery promise and returns policy so agents can quote and buy directly. Winners become the default answer; losers become invisible.
"Something for an October wedding, I'm between sizes" becomes a real shortlist, using fit history and return reasons. Attacks the two biggest costs in the category at once: abandonment and returns.
An agent can always find cheaper. It cannot find nearer. Collection on a route they already drive, fitting, alterations and returns without a courier are things only an estate offers. AI makes the visit personal: the colleague knows what their agent shortlisted and has it ready.
200,000 SKUs described, attributed, localised and kept accurate in hours, not quarters. Attribute quality has quietly become a distribution asset rather than a back-office chore.
AI reads the photo, the order, the courier scan and the policy, then resolves (refund, replace or reject) before the customer chases. Returns stop being pure cost and start creating promoters.
The customer tells their assistant to buy and nothing gets in the way: identity verified, spend limits honoured, no fraud rule mistaking a good agent for a bot. Where agent purchases simply complete, agents come back, and you set the terms everyone else inherits.
Build offers that survive comparison. Publish atomic prices an assistant can select, and the non-price terms that decide a close call: delivery, customisation, warranty, returns. Bundles now have to beat their own parts, so know who was subsidising yours.
AI reads weather, local events, social signal and live sell-through to move stock before the gap appears. Availability is the one thing an agent can't be persuaded around and the one that fewest lost sales get blamed on.
Agentic commerce runs both ways. Your agent makes the offer, to the customers worth keeping and only as far as the margin allows, and their agent judges whether it genuinely serves them. Retention becomes a negotiation, and a customer's agent cannot be flattered.
Strategy and design support for an industry first agent-supported grocery shopping service
View the case study →
Running a hackweek and AI training to drive faster adoption and help accelerate product delivery 3x
Identifying and designing parts of a complex B2B sales journey which can be automated with AI tooling
View the case study →
Delivering new eCommerce experiences and developing new product and design capabilities
View the case study →
[Bio]
[Bio]
[Bio]
[Bio]
We challenge you and your team on how AI is changing your customers' world, and your competitors' operations.
Map the impacts of AI on your proposition and prototype the future with us.
Use our AI tooling to accelerate delivery of your priority initiatives.
