Barkly, as Google and the AI engines see it.

Barkly's customer is a woman who walks her dog alone. In 2026 she increasingly asks an AI engine before she ever types into Google. So we asked her question, word for word, and read what came back.

This is not a reputation problem. It is a presence problem, and the two have very different fixes. The engines recommend what their trusted sources have written about. She's Birdie appears in Forbes, Wired and the big safety-gear listicles. Barkly does not, yet. Section 05 shows exactly which doors to knock on.

Then we played the other scenario: she saw a Barkly ad and checks the brand before buying. "I saw an ad for the Barkly dog leash. What is Barkly and is their leash any good?" Here the picture flips. All engines answered correctly and positively: the veterinarian co-founder, the 130 dB alarm, the strobe, honest pros and cons, a "worth considering" verdict. Your two independent reviews, Vetstreet and Petful, are doing almost all of this work.

Three details the engines surfaced that are worth acting on:

• ChatGPT flagged that your listings contradict each other on sizing ("up to ~44 lbs" on one, "up to 100 lbs" in a review) and said it would "trust the current retailer listing before buying". It even told our shopper the Chewy listing was out of stock, while Google's crawl of that same listing shows it in stock. Whichever side is stale: one master product sheet, synced across the site, Amazon and Chewy, ends the contradictions the engines repeat out loud.
• Claude qualified its praise: the reviews are "mostly from pet-review sites that were given the product", with little long-term independent feedback yet. The engines can tell seeded coverage from earned coverage, which is exactly why Fix 02 matters.
• The distribution itself looks strong: Amazon, Chewy, Walmart and eBay listings all surfaced, and your press release traveled through the newswire network.

The pattern matches Section 01 exactly: whoever wins the gateway articles wins the category, in classic search and in AI answers alike. Barkly wins where no one else plays, and is invisible where the volume is.

The deepest root cause is almost funny: the Shopify store name is set to the full tagline instead of "Barkly". Machines read it everywhere:

// Organization schema (how you introduce yourself to Google's Knowledge Graph)
{
  "@type": "Organization",
  "name": "The Barkly Dog Leash With a Built-in Safety Alarm & Strobe Light",
  "sameAs": [ null, null, null, null, null, null, null, null, null,
              null, null, null, null, null, null, null, null ],
  "url": "https://mybarkly.com"
}

// Product page, the review schema. 117 reviews exist; none of them count:
"aggregateRating": {
  "ratingValue": { "error": "json not allowed for this object" },
  "reviewCount": 117
}

That is how Barkly introduces itself to Google and to every AI crawler: a 60-character name, 17 empty slots where Instagram, TikTok, Amazon and press profiles should be, and a review score replaced by an error message. It explains a lot of Section 03.

• Homepage title is doubled: the tagline appears twice (page title + store name), ~140 characters, truncated in every result.
• Your review schema is broken by a template bug: the product page declares 117 reviews, but the rating value literally contains an error string instead of the score (exhibit above). Google cannot award stars to an error message, and the engines get no review signal.
• robots.txt was reduced to two lines which block your /collections pages and dropped the default "Sitemap:" line in the edit. With one hero product this costs little today (your homepage does the selling), but the sitemap pointer and the default protections are free to restore, and collections will matter the day the range grows.
• Empty H1 on the homepage (the logo sits where your main heading should be) and a 255-character meta description (cut at ~160).
• 3 blog posts total: the "solo walk safety" territory you should own has almost no content on your own domain.
• Good news: Shopify's new agents.md for AI shopping agents is live on your store, so the plumbing for agentic commerce is in place. Lighthouse's new agentic-browsing audit still scores the homepage 67/100 (images missing alt text, malformed accessibility tree): worth cleaning as AI shopping agents become real traffic.

None of this is dramatic. All of it is fixable in hours, and most of it traces back to one settings field.