Others read your code and guess. Every Sunday our agents try to buy from retailers across Australia, the UK and the US, and we record where each one stalls. Your two scores: NOW, how well you sell to a person today, and NEXT, how well you'd sell when the shopper is an AI agent.
Want the full picture? The $495 report sends an agent through your checkout, ranks you in your category and lists the fixes in order.
Say a shopper asks their AI assistant for a mid-range espresso machine under $600, delivered by Friday. The assistant doesn't browse. It reads product data, checks price and stock, tries to complete the order, and picks whichever store made that easiest.
Your NOW score is what you already track: findability, trust, how smoothly a person can buy. Your NEXT score is whether that assistant can do the same job without a human. Most retailers score well on the first and poorly on the second. The width of that gap tells you how much of your future sales an agent will hand to someone else.
Real example, electronics, Australia.
An AI shopping agent visits your site live, the same way a shopper's assistant will. It reads your product data, tries your checkout, and weighs your reputation.
One method, two scores: NOW for how well you sell to a person today, NEXT for how well you'd sell to an agent, weighted for your category.
The report lands in your inbox: factor-by-factor scores, your rank in your category, and a 90-day fix list ordered by impact. PDF within 24 hours. No calls.
Trust (weight 30, heaviest), Discoverability (25), Suitability (25), Transactability (20). The signals a person weighs when an AI surfaces options and they decide: reviews, search presence, site quality, delivery fit.
Agentic (weight 30), Technical (30), Market (20), Sentiment (20). What an AI shopping agent needs: a checkout it can complete, product data it can read, citations in AI answers, a reputation it can parse.
A posted commodity (books, groceries) meets the agent future fast; an anchored, touch-test purchase (jewellery, furniture) stays human-led longer. Tier X · mixed-velocity spans both, a candle shifts to agent-buying long before a kitchen does. Tiers tune the weights to your category's velocity; they modulate the score, never excuse it.
Sean Howell led AI adoption inside a global retailer before building blynk. Every score here comes from a method he'd have wanted when he was the one being asked "are we ready?" in the boardroom.
Read the essays →
