The agent does not have a favourite.

A person's trust is a shortcut. An agent's is a checklist.

Stock count, current price, delivery time and returns terms decide how a product comparison goes. A shopper skips most of that check when they already trust the brand. An agent skips none of it, for either option, every single time.

When someone has bought the same candle three times running, the fourth purchase barely gets a second look. That is not laziness, it is a reasonable shortcut built from a real track record. A shopper wants a birthday gift, a cinnamon candle around £24, ships this week, and asks their agent to find one. The agent does not know or care that the shopper's aunt has bought from the same small candle maker in Bristol for years. It checks the Bristol candle maker's page against a national homeware chain's page, on the same four facts, at the same moment, and picks whichever one clears the bar.

There is no accumulated goodwill sitting in the agent's memory the way there is in a returning customer's. Every comparison starts from nothing, which sounds harsh until you notice what it removes: the years a smaller name spends being overlooked simply for being less familiar to more people.

That changes who gets a fair look.

For the small brand, this is mostly good news. A household name's recognition counts for nothing extra here, its size does not win it anything the page itself does not back up. If the national chain's listing says "in stock" but the warehouse feed has not been touched in three days, and the Bristol candle maker's own site has the honest stock count from this morning, the smaller business wins the comparison on the same four facts, not on being memorable.

The cost of getting this wrong lands on whoever has the stale page, not on whoever is smaller. A shopper whose agent orders from the national chain, only to get a cancellation email two days later because the item was never really in stock, has lost time the smaller, honestly updated competitor would never have cost them.

It cuts both ways.

None of this means a big name automatically loses. A well-run national chain with an accurate, live-synced feed still wins fairly against a smaller competitor whose page has gone stale, because the same rule that removes an unearned advantage also removes an unearned disadvantage. Size stops being the variable that matters at all, in either direction, once the comparison is running. What decides it is which page was kept current, and a business with more resources to keep its feed accurate has every reason to expect that to show up in its favour.

It also raises the bar on being checkable.

The other side of having no favourite is that the agent also extends no benefit of the doubt. A person forgives a slow-loading returns page because they already trust the brand from past experience. An agent checking that same page for the first time treats a vague returns policy exactly like a missing one, it cannot verify what is not there, so it moves on to the option it can.

A business that has been riding on reputation for the parts of the page it never quite finished writing properly will find that reputation buys nothing here. The four facts either hold up today or they do not, and "usually accurate" is not a category an agent can act on.

Those four facts, stock, price, delivery and terms, need to stay live and checkable from a business's own systems, not sat on a static page that goes stale the moment nobody updates it. That's what Selfe builds, so a small brand's honest "in stock, ships tomorrow" beats a bigger name's outdated one on the facts, not on name recognition.

Does a well-known brand still have an advantage with AI shopping tools?

Not in the comparison itself. Recognition might get a brand mentioned more often in a general answer, but a purchase decision checks the same specific facts for every option, regardless of size.

What facts matter most for this kind of comparison?

Current stock, an accurate price, a real delivery estimate and clear returns terms. Whichever page has all four, current and specific, tends to win the comparison.