AI visibility for Shopify stores: a 30-day plan

A Shopify store handles some of the plumbing itself, such as sitemaps, a structured data filter and a catalog feed to AI channels, so the work is mostly checking and filling gaps, not engineering. The risk is the opposite one, spending a month on tasks whose effect you cannot see. This plan orders the work so the cheap, documented steps come first and the uncertain ones come after a baseline.

Hours below are planning estimates by Dopbot for one person and a store with up to 25 products, not measurements. Scale the product-facts stage by your catalog size. Shopify features are described as of October 2026 and admin screens change.

What does the 30 days look like?

That sums to 26 hours. If you have only an afternoon, do the access stage and the product facts for your five best sellers, and skip the rest until you have time to measure.

Days 1 to 3: how do you set a baseline?

Write four questions that name no brand: best overall, a budget with a price ceiling, premium, and a specific use. Use two assistants, for example ChatGPT and Perplexity. Ask each of the four questions three times on each, in clean sessions. That is 24 answers, 12 per assistant. Save the full text and every source list.

Then count, per assistant, the answers that recommend your brand, and write the total beside the count: "3 of 12", not "25%". Also list the domains in the source lists. You will use them on day 22.

Why keep the assistants apart? In the October 2026 edition, excluding the alternative-to-a-brand question, the brand lists of OpenAI's API with web search (gpt-6-luna) and Perplexity's Agent API with the Sonar model (perplexity/sonar) overlapped by 47.0% on average per category (brands both recommended, divided by brands either recommended), and answers recommended 2.3 brands on average. One assistant's result says little about another's, and a shortlist has few places. The manual measurement method has a spreadsheet layout; it uses five question types asked four times, and the layout carries over.

Do not read too much into a zero. If your brand appears in none of 24 answers, a 95% Wilson interval still runs up to about 14%. These intervals treat the 24 answers as independent draws, but they are three repeats of four questions on two assistants, so the real uncertainty is wider. Per assistant (12 answers) a zero still runs up to about 24%. The baseline tells you where you stand roughly, not exactly.

Days 4 to 7: can the bots and Shopify reach your products?

Estimate: 2 hours.

1. Read your robots.txt. Open yourstore.com/robots.txt. Shopify says its default allows crawlers to request your public content and disallows admin and cart pages. Search the file for Disallow: / under User-agent: *, and for lines naming OAI-SearchBot or PerplexityBot. OpenAI and Perplexity both recommend allowing their search bots if you want to appear in their search results.
2. Find out whether your theme has a custom robots.txt.liquid. Shopify says that template replaces its generated robots.txt output. Before you change anything, save a copy of the current output, as that page advises. Shopify's help center labels editing it an unsupported customization and warns that incorrect use can result in loss of all traffic.
3. Check that private mode is off. Shopify says private mode hides all of your online store's pages from visitors and search engines, and its Catalog requirements say a store in private mode is ineligible. Shopify's steps place the Private mode setting in the Store access section under Online Store, Preferences.
4. Check how Markets handles crawlers. Shopify says that for markets sharing a domain, content is not indexed by search engines. Shopify does not say which market's version a crawler sees, so treat your default market's version as the one a bot may read and keep it accurate. Shopify's page says content on markets with dedicated domains is indexed, so check which setup each market uses.
5. Open your sitemap. Shopify generates sitemap files automatically at your domain's /sitemap.xml and updates them when you add pages, products or collections. Confirm it loads and lists your product pages.
6. Review any app that blocks countries or bots. Open its settings and read what it does to search bots. Shopify says its own Cloudflare protection is always active, and that it does not support adding your own Cloudflare proxy. So on a standard store, the controls that affect search bots, as far as Shopify's documentation shows, are robots.txt, apps, redirects and private mode.

A change to robots.txt can take about 24 hours to be reflected, according to OpenAI's bot documentation. Perplexity says its systems may take up to 24 hours to reflect changes.

Days 8 to 14: which product facts need fixing?

Estimate: 8 hours, at about 20 minutes a product.

For each product, in this order:

1. Set the product category. Shopify says all of your products should be assigned to a specific category in its Standard Product Taxonomy.
2. Make the title say what it is. Brand, product type, model and the variant that matters, in the words a buyer uses.
3. Write the facts into the description as text. Price-relevant details, dimensions, materials, what is included, who it suits and who it does not, and care and warranty terms. The product page audit is a checklist for this.
4. Add alt text to every image. Shopify says alt text is an important part of a product description and that 125 characters or fewer is recommended.
5. Complete your store policies. Shopify's ChatGPT channel requirements include terms of service, privacy and return and refund policies completed in Settings.
6. Put required legal disclosures early. The same page advises that descriptions include any relevant legal disclosures in the first 6,000 characters for the ChatGPT channel.

Fix your five best sellers first, then continue in order of revenue.

Days 15 to 21: what are your markup and feeds doing?

Estimate: 4 hours.

• Check the Product markup. Shopify's structured_data filter converts a product into schema.org markup: a Product when it has no variants and a ProductGroup when it has one or more. Your theme has to call it. View a product page's source, search for application/ld+json, and run the URL through Google's Rich Results Test. The structured data guide covers what to check.
• Add Organization markup. Google recommends placing it on your home page. It lists a return policy and a shipping policy among the Organization properties you can add.
• Check the Google and YouTube channel. Shopify says it automatically syncs your products and relevant store information with Google Merchant Center.
• Open Sales channels, then Agentic. Shopify says agentic storefronts are active by default for eligible stores, and that products are made available to AI channels through Shopify Catalog. Catalog needs a title, at least one image, a price above zero, and publication to your online store. The ChatGPT channel requires that you sell to customers in the United States.
• Read the feed guide. The product feeds post covers Merchant Center, ChatGPT and Shopify in detail.

Do not hand-edit markup you do not understand. If a product page's source holds two Product blocks, find where each comes from, your theme or an app, and keep one.

Days 22 to 28: what about reviews and outside sources?

Estimate: 6 hours.

• Ask every buyer for a review. Not only the happy ones. The FTC's final rule bars businesses from offering incentives conditioned on reviews expressing a particular sentiment.
• Show reviews on the page and mark up only what is visible. Google's rules forbid marking up content that readers cannot see on the page.
• Work from your day 1 source lists. Note which review sites and retailers appear for your questions, on each assistant. Check that your listings there are accurate, and read each review site's published method before pitching it.

In the October 2026 edition, a domain on Dopbot's hand-made retailer list was in 76.7% of Perplexity's 787 retrieved source lists (27.3 URLs per answer on average) and in 17.6% of OpenAI's 800 answers' linked pages (2.2 per answer). The lists are different kinds and lengths, so the two shares are not comparable. A domain in a list does not show that it changed a recommendation.

Days 29 and 30: how do you re-measure?

Estimate: 3 hours. Ask the same four questions the same way, three times on each assistant.

Compare counts with an interval. A brand named in 6 of 24 answers has a 95% Wilson interval of about 12% to 45%. The methodology says changes will be shown only after an exact test between two updates of the same size gives p below 0.01, and notes that this test is not built yet. As a planning check, Fisher's exact test (two-sided) would need 16 of 24 against a 6 of 24 baseline to reach p below 0.01. A move from 6 to 9 is well inside the noise. The post How many AI answers you need before trusting a change explains the math.

Two more cautions. You changed several things at once, so even a real change cannot be credited to one of them. And 30 days is short: model and search changes happen on their own schedule. Treat this plan as removing blockers and fixing facts, which is worth doing whether or not answers move this month, not as an experiment.

If you want a free first look at how often assistants recommend your brand before you start, the snapshot gives one.

What this plan does not tell you

• The October 2026 numbers are from APIs, not the consumer apps, and the apps can differ.
• No source we know of shows that these steps raise recommendations. The access, policy and catalog steps are documented requirements; the rest are reasoned.
• Shopify's programs have requirements that differ by channel. The ChatGPT channel requires sales to US customers, and Catalog requires the Starter plan or higher, except for stores on the Agentic plan that have not yet published to a paid plan.

Frequently asked questions

Do I need a developer for this plan?

Mostly no. The product facts, store policies, alt text and channel checks are admin tasks. Adding Organization markup or editing robots.txt.liquid means editing theme code, which Shopify labels unsupported for robots.txt, so involve a developer for those.

Should I allow every AI bot in robots.txt?

Not as a blanket rule. Search bots, training crawlers and user-initiated fetchers have different documented roles, and OpenAI says each setting is independent of the others. Decide per bot, and write down why.