Three questions every business should ask about its AI readiness.

Three questions, asked in order

A homeware retailer wants to know if it's "AI ready." The honest answer isn't a single number, it's three separate questions, asked in the order an agent encounters them: can it find the business, can it establish what's true, and can it make progress toward an actual purchase. A review that skips straight to a score without answering these in order is measuring something, just not readiness. This site's own piece on the readiness score explained covers how those same three questions show up once they're compressed back into a single number; this one walks through them directly, before any score gets involved.

Can an agent find the right starting point?

An agent cannot select a business it cannot find or identify. Check whether the business name, offer, location and basic contact route are clear enough for a customer request to reach the right place. For the homeware retailer, that means a product page that states plainly what it sells, where it ships to, and how to reach a person if something goes wrong, not buried in a contact form three clicks deep.

The bar here is narrower than ranking for every phrase: whether the business can be recognised and considered at all when an agent is doing the research on a customer's behalf.

Can it establish what is true?

The next question is whether the information that changes a choice is available, current and understandable. That includes the offer, limits, policies, price conditions and the detail a customer would reasonably need before proceeding. A returns window stated once, clearly, on the product page clears this bar. The same policy split across three different pages, each slightly different, does not.

Where a fact changes, identify the system or person that owns it. A readiness review should expose uncertainty rather than dress it up as an answer.

Can it make useful progress?

Useful progress may be a confirmed option, a qualified quote, a booking request or a clean hand-off. It does not always mean automatic payment. The business decides which actions are permitted and which need a person or customer approval. For the retailer, that might mean confirming stock and price, then handing the actual checkout to its own site rather than trying to complete the purchase itself.

The practical test is whether the context gathered so far can travel to that next step without forcing the customer to begin again from nothing.

Turn the answer into a next move

The point of the review is a priority, not a verdict. Find the first break in a real route, name the fact or hand-off that would remove it, then make that one change measurable. That is how readiness becomes commercial progress rather than a score nobody acts on.

Selfe runs these three questions against a business's real, connected systems, rather than against a static description of them, finding the first actual break in the route, not a hypothetical one.

What counts as "useful progress" if a business doesn't want to sell directly through an agent?

A confirmed option, a qualified quote, or a clean hand-off to the business's own checkout all count. Useful progress doesn't require letting an agent complete the purchase itself.

Does a high readiness score mean nothing is worth improving?

No. A single first fix, the one that removes the next real dead end, usually does more than chasing a marginally higher score everywhere else at once.