Your Customer Already Has An Advisor...They Need Something Else From You

Shoppers now arrive to retail units much more informed. They get this information from a machine and once they step in to your world they want something more. The question is: have you designed your unit and trained your frontline to offer that 'something more'?

State of e-commerce report 2026 shows that AI-referred traffic to retail sites rose 805% year on year over Black Friday, and AI platforms are on course to drive close to 21 billion dollars of retail spending this year. Most of the commentary asks whether the machines will replace the physical units. The more useful question is what happens in the twenty seconds after someone walks through your door already knowing what they want.

The Gap Just Widened

86% of shoppers begin their research online. That part is old news. What has changed is that the research is now done for them, by something that never gets tired, never loses interest halfway through, and never says “let me just check with my colleague.”

38% of consumers already use AI when they shop, and 80% expect to use it more. AI-driven orders grew fifteenfold in a single year.

Now put that in your store. 83% of shoppers believe they know more about the product than the person serving them. 79% still say a knowledgeable associate matters to them. Both things are true at once, which is the odd bit.

So, your customer arrives holding a comparison your team has not seen, from a source your team cannot check, about a range your team was briefed on in a fifteen-minute huddle before opening. Your team will not lose an argument about specifications. They will lose an argument about preparation, and they will lose it at the cost of the business.

The Four Jobs

If the machine has taken information, what is left? Four things, and every one of them is harder than reciting a spec sheet.

Know enough to correct AI's findings: AI is fluent, fast and occasionally wrong, and it is wrong with total confidence. Your team should know the five claims it gets wrong about your category and be able to correct them politely, with evidence, without making the customer feel foolish for believing it. Write those five down. That is a training session, not a strategy day.

Narrow down the options for the shopper: The machine hands over options. It does not hand over permission. A shopper standing in front of three near-identical products does not need a fourth suggestion. She needs someone to say “this one, because of what you just told me about how you’ll use it.” That is judgement, and it does not come out of a database.

Use the power of experiencing products in person: Weight, fit, noise, finish, how the drawer feels on the third pull. Nothing online closes this gap. If the products where touch decides the sale are sealed in boxes on a high shelf, you have hidden your only genuine advantage.

Vouch for what your business can do that AI cannot: An AI cannot stand behind anything. It will not process the return, chase the warranty or remember the customer’s name in March. Your store can do all three, and that promise is worth more now than it was when you were merely the place that had the stock.

The Catch

Do not send your team out with a script about AI being unreliable. Customers have used it, it mostly worked, and telling them otherwise makes you sound frightened. Position the store as anti-AI and you have picked a fight with your own customer’s judgement, which is a fight nobody wins. Correct specific errors, quietly.

The Practical Reset

Four questions. Answer them this week, with your team in the room.

• Ask an AI assistant about your best-selling product, right now, on your phone. Would your Saturday team recognise that answer as accurate? If not, you have just found this week’s training and it took ninety seconds.

• When a customer opens with “I read that…”, what does your team say next? If the honest answer is silence, or “I’m not sure”, that is your whole problem in five words, and it is not the customer’s fault.

• Which of your products cannot be judged without being touched? Those are your reason for existing. Now go and look at where they are sitting on the shop floor.

• What can you promise that no AI can? Say it out loud to another human being. If it takes you more than ten seconds to answer, assume your customers worked that out some time ago.

The Part The Machine Cannot Do

The machine has taken the answers. It has not taken the judgement, the hands, or the promise. Those were always the job. We simply got away with pretending otherwise back when the customer knew less than we did.