Blog / Industries

AI visibility for ecommerce brands: get recommended

The questions shoppers ask AI assistants before they buy, how to see whether your products are named, and the pages and places that get you recommended.

By the BotSay team21 Oct 20266 min read

On this page
  1. Where shoppers meet AI answers
  2. The questions to track
  3. What a good answer looks like for you
  4. Pages the engines lean on
  5. Places to be mentioned
  6. Reading the change week to week

A shopper who wants running shoes, or a gift for a keen cook, can ask an AI assistant before they open a single shop. The answer names a few brands, says a line about each, and often links to the pages it used. If you sell online, that answer is a shop window you do not own, and it pays to know whether your products are in it.

Where shoppers meet AI answers

Shoppers put these questions to ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews and Google AI Mode. Some want a quick shortlist. Others describe their situation in detail: "a waterproof jacket for walking the dog in winter, under a set budget, that still looks smart in town".

The answer to a question like that usually names a few options in order. A brand named first gets read first. A brand left out has to win the shopper later, against everyone the assistant already recommended.

The questions to track

We sort shopper questions into three kinds. Write them the way a shopper would, with the details they give: who it is for, the budget, the size of their home, where they live.

Category questions

The shopper knows what they want to buy and has no brand in mind.

These show whether the engines place you in your category at all. Give them the largest share of your questions.

Need and situation questions

The shopper describes a problem, a person or an occasion.

The engines answer with advice and name products along the way. These questions are where a smaller brand with a clear specialism can be named next to much bigger ones.

Comparison questions

The shopper has names and wants to choose.

Track the ones that name a rival and leave you out: they show whether the engines offer you in the rival's place. A question that names your brand nearly always gets you named, so BotSay counts the five measures over the questions that do not name you.

Give each product line or category you want to win at least ten questions. Our guide How many questions do you need to track AI visibility? shows how the number you track sets how precise your figures are. Add your questions one by one, or accept the ones BotSay recommends from your website address, and edit them as you go.

If you sell in one country, ask as a shopper there would, since prices, delivery and stock differ by market. BotSay asks each question as its own country on every engine that can be pointed at one: the United States, or the country you choose for that question. Gemini cannot be pointed at a country, so it answers the same wherever it is asked from.

What a good answer looks like for you

Here is the kind of answer you are aiming for: a buyer question, a short list, and your brand named first.

ChatGPT, asked this week

What is the best waterproof jacket for walking the dog in winter?

For winter walks, Fernhollow is a strong choice: it stays dry in heavy rain and comes in long and short fits. Tidelark suits hill walkers, and Marrowby is worth a look if you want a jacket that looks smart in town.

You are named first

Example, using an invented brand

Five measures describe where you stand: Mention rate, Share of voice, Average rank, Your pages cited and Sentiment. For a shop, start with two. Mention rate is the share of answers that name you, and Average rank is how early you come when you are named, where lower is better. Together they tell you whether you make the list, and how near the top.

Then read the sentence each engine writes about you. "A good choice for wide feet" sends a different shopper your way from "a budget option". If an engine describes a range you no longer sell, the page it took that from is often one you can update. BotSay sorts the answers that name you with the least favourable first, so the ones worth fixing come to the top.

Pages the engines lean on

Engines that search the web write their answers from pages, and link to many of them. Your pages cited is the share of those links that point at your own site.

In ecommerce, the pages that come up tend to be:

The example below sorts one week's cited pages by owner: the brand's own, its rivals', the places it could be listed, and everyone else.

Pages the engines cited this week, by owner
  • Your pagesYou
    such as fernhollow.example/dog-walking    17.2%
  • Your rivals' pages
    such as tidelark.example/hill-walking    31.5%
  • Places you could be listed Get listed
    such as outdoorgearreviews.example    38.9%
  • Everyone else
    such as walkingnews.example    12.4%

Example, using an invented brand

To make your own pages easier to cite, answer the shopper's question in plain words near the top, with specifics an engine can repeat: materials, sizes, delivery, returns, the price. Write a buying guide for each question where rivals are named and you are not. Keep pages open to the crawlers the engines use, and keep them current, because an engine that cites an old page repeats an old fact.

BotSay lists the pages the answers cited, and once you confirm which sites are yours and your rivals', it sorts each one as yours, a rival's, or somewhere you could be. For every answer, it also shows the searches the engine ran to write it, so you can see the words a page of yours has to answer.

Places to be mentioned

Many pages the engines trust are on other people's sites, and you can often get onto them.

BotSay lists the pages the engines read or linked in answers that named a rival, each with a label for what to do, such as Pitch them or Claim your profile, and whether your name is already on the page. Start at the top.

Reading the change week to week

Ask the same questions in the same words every week. A Mention rate counted from a handful of answers can swing on a single reply, while one from a few hundred holds steadier. BotSay shows Mention rate with its likely range and the number of answers behind it, and calls a change real only once it is larger than chance would explain.

Three product lines of ten questions on six engines is 180 answers a week. BotSay asks your questions on all six engines every week, through each provider's official developer access or a licensed search-data provider, and works out the five measures for you, with your rivals beside you. Your reading arrives in the weekly email.

Then pick the category where rivals are named and you are not. Publish the buying guide or fix the product page, and mark the day it goes live in BotSay. It compares the weeks before that day with the weeks after, on the questions the page was meant to win, so you can see whether it paid off.

What you gain with BotSay

  • Your shopper questions asked on six engines every week, with the answers read and counted for you.
  • A Mention rate you can report with confidence, with its likely range and a label when a change is real.
  • Every page the engines cited, sorted as yours, a rival's or somewhere you could be, and a list of the places to pitch, claim or join.
  • Pages to write, ranked by how many questions each could win, and places to be mentioned, ranked by how often the engines lean on them.
  • One fixed monthly price with everything included and no add-ons. Plans differ only in the number of questions, from $59 a month for 60 questions.

Keep reading

Your first reading runs right after setup

Find the page to write next.

Monthly plans from $59. Cancel at any time.

Get started