Thoughts / Gislain Armand
Earning the AI Recommendation
On how AI agents interpret your brand, shape your customer’s choices, and change what it takes to earn their business.
For years, resellers have invested in being found. Understand the search terms, improve your visibility, get the customer to your website. Then explain why they should buy from you.
Now consider what happens when that customer gives the research to an AI assistant. They describe what they need, what they can spend, and what matters to them. The assistant can search, interpret the available information, and recommend options before the customer ever visits your store.
Part of the selling happens before you get the conversation.
In From Search Queries to AI Prompts, I wrote about the advantage of understanding the machine. This is the next consequence for our industry: we have to build businesses whose value can be understood by the systems helping our customers choose.
The New Selection Layer
Search optimization still matters. Google’s guidance is clear that its AI search features rely on the Search index and core ranking systems. The work of making useful information discoverable remains relevant.
But being found does not settle the buying decision. An assistant may also interpret whether a product fits the customer’s requirements, which trade-offs matter, and which seller offers suitable terms. OpenAI’s shopping documentation, for example, distinguishes selecting products from ranking the merchants selling them, and describes how the request and available context can influence the results.
That distinction matters to a reseller. A manufacturer’s product can be recommended without your business being the recommended place to buy it.
There is no single universal AI index or ranking to master. Different assistants use different sources and methods. They may have incomplete information, and their answers can be wrong. No placement is guaranteed. The strategic response is to make your business easier to evaluate accurately, while continuing to earn the customer’s trust directly.
Same Product, Different Recommendation
Imagine a customer considering a pair of Sharp headphones alongside models from other brands. They ask their assistant to help them choose. What would make the Sharp option the right recommendation?
The answer starts with the customer. Someone who spends their mornings on a crowded train may care about comfort, noise control, and taking calls. Someone buying headphones for exercise may place more weight on a secure fit and suitability for sweat or outdoor use. A listener who spends evenings enjoying music may be willing to pay more for the sound and listening experience they prefer.
Those are different buying decisions. The same Sharp model could be a strong candidate in one and a poor fit in another. We would have to examine the actual model and the alternatives before drawing that conclusion.
| Customer’s situation | What the comparison needs to establish |
|---|---|
| Commuting and calls | Relevant noise-control features, microphone performance in realistic conditions, device compatibility, and evidence about comfort over time. |
| Exercise and outdoor use | Fit and stability, any documented water-resistance rating and its limits, and whether the design suits the intended activity. |
| Music and longer listening | Credible listening evaluations, fit and comfort, and the trade-offs that matter within the customer’s budget. |
Hypothetical buying situations, not claims about any Sharp model. Features and suitability must be verified for each product being compared.
If the Sharp model meets the customer’s important requirements at a suitable total price, while a more expensive alternative mainly adds features they do not need, there is a clear reason to consider Sharp. If another model better addresses a requirement that matters, the recommendation can change. A useful comparison explains that trade-off instead of declaring one brand universally superior.
Brand also carries meaning that a specification sheet cannot fully capture. One customer may already trust Sharp because of products they have owned. Another may prefer a particular audio brand, appearance, or design. When the customer expresses those preferences, they become part of the decision. An assistant can take them into account without experiencing that attachment itself.
This creates an opportunity for a reseller who understands the audience. Repeating “premium sound” tells us very little. Explaining which customer a particular model suits, what supports that judgment, and where it falls short gives the comparison substance. Manufacturer documentation can establish supported features. Credible testing can help assess performance. Comfort and sound preferences still leave room for personal judgment.
Then comes the second decision: where to buy. Suppose two stores carry the same Sharp headphones. One offers a lower total price. The other, in this example, provides clear compatibility advice, a delivery estimate that fits the customer’s deadline, and plainly explained support and return conditions. Those services have to be real, and any limits on returning opened headphones need to be visible.
A buyer who already knows exactly what they want may choose the lower price. A buyer uncertain about fit, or buying before a trip, may value the other offer more. The reseller earns consideration by making a meaningful difference for that customer and providing evidence an assistant can evaluate.
You do not need to be the right business for everyone. You need to make it clear whom you are the right business for.
Your Brand Has to Mean Something
This is where brand becomes a leadership issue. If your identity is “we sell everything at great prices,” you have given the customer very little reason to prefer you. A specialist who understands a particular customer’s problems has something more substantial to explain.
Building that identity requires choices. Which customers do we understand particularly well? What do they struggle with? What can they reasonably expect from us that makes choosing our business worthwhile? Those answers should influence the products we carry, the advice we give, and the service we are prepared to provide.
Consider two hypothetical audio retailers. One builds its business around broad selection and low prices for buyers who already know what they want. The other helps people choose headphones for everyday work and commuting, with a carefully selected range and knowledgeable assistance. Both may carry the same Sharp model. Both can offer value. Their brands mean different things because they have made different choices about whom they serve and how.
The second retailer has to demonstrate that judgment. It should be able to explain why a product belongs in its range, which situations it suits, and when another choice would be more appropriate. A recommendation against the more expensive option, when that option adds little for the buyer, can show the customer what the business stands for. That is a more substantial expression of expertise than adding “trusted experts” to every page.
A manufacturer’s name may bring a product into consideration. The reseller still needs a reason to be chosen. Your visual identity and tone help people recognize you, but the distinction must survive beyond them. If your logo disappeared from a product page, would the selection, explanation, and offer still reveal something particular about your business?
Your brand should make your expertise recognizable. Your offer should show what it means in practice. Specific explanations, clearly stated service responsibilities, and genuine customer feedback provide evidence people can assess. Where an assistant encounters that information, it may also help it understand the business. None of those elements is a universal AI ranking formula, and repeating a claim does not establish that it is true.
An AI agent is not a human customer. It does not experience your service or develop loyalty to your team. It works from the information and instructions available to it. The person buying still lives with the result. Building a brand solely to impress the software would miss the reason the brand exists.
The strongest position is to serve a defined customer well and make the evidence of that work easy to understand. A clear brand gives the customer a reason to prefer you and gives an assistant something meaningful to explain.
The Business Behind the Promise
Clear communication exposes the quality of the business behind it. A delivery promise is only as good as the inventory and fulfillment supporting it. Product expertise depends on reliable manufacturer information. A support commitment requires someone capable of honoring it.
Return to the customer buying headphones before a trip. A useful delivery estimate depends on the exact model being available, the order reaching fulfillment before the relevant cutoff, and the delivery service reaching that destination in time. “Fast shipping” hides those dependencies. Leadership has to decide what the business can responsibly promise and what happens when the information is uncertain.
This is where technology becomes commercially important. Connecting inventory and product information can reduce the work of keeping an offer current. But a connection still needs dependable source data, checks for failed updates, and someone responsible when the information is wrong. A polished product page does little for a customer whose order cannot be fulfilled as described.
The same discipline applies after the sale. If we promise help choosing or setting up a product, who provides it, and when are they available? If the product develops a fault, what does the reseller handle and what belongs to the manufacturer? If the applicable change-of-mind return policy excludes opened headphones, that limit belongs in the offer before the customer buys. The answers should remain consistent across the product page, checkout, and support conversation.
These choices also have to make commercial sense. Advice, delivery options, and after-sale support require resources. A reseller should understand the cost of the service it promises and whether its pricing and supplier arrangements can sustain it. A useful distinction that the business cannot afford to maintain is a fragile foundation for a brand.
That is why I see this as a leadership responsibility. Marketing can explain the promise. Technology can keep the information current. Operations and supplier relationships determine whether we can deliver. Those parts of the business have to agree, and someone has to own the complete customer outcome when responsibility crosses between them.
The evidence we collect should help us improve that outcome. Compare delivery estimates with actual arrivals. Examine why customers ask for help, return products, or say an offer was unclear. When a promise repeatedly fails, correct the process or narrow the promise. Publishing more persuasive copy will not repair the underlying problem.
We can also ask an assistant to evaluate the offer using realistic customer situations, then check its explanation against the facts. Does it identify the right product, understand the service, and recognize the limits? That exercise can reveal unclear information or an error in the assistant’s answer. A favorable response is still only one observation, not proof of reliable placement or future recommendations.
For resellers, understanding the machine should bring us back to a fundamental commercial question: why should this particular customer buy from us? Choose the customer you intend to serve, build an offer that addresses their actual constraints, and make sure the business can deliver what the offer says.
If your next customer asked an AI assistant why they should buy from you, what evidence would it have to work with?
