The 5 Trust Signals AI Overviews Demand to Cite Your Brand — the Freshpet Case
Summarize this article with AI
A client calls me one Tuesday morning. He invested $8,000.
He sells joint supplements for dogs. $8,000 in editorial SEO. 14 months of work. Dozens of pages in the top 3.
And yet.
When I search « best dog joint supplements » in Google, an AI Overview appears. It cites three brands. His isn’t there. Not once across the 47 queries in his market.
The problem wasn’t his content. It was the absence of signals that language models use to decide who’s a trustworthy brand.
« Rankings prove relevance to a query. Citations require more: trust signals the system can verify about the brand itself. » — Intero Digital, Webinar Search Engine Journal, August 2026
Freshpet hit exactly the same wall. An established brand, a catalog of fresh food for dogs and cats, solid SEO. And still zero citations in AI Overviews. Their digital team, led by Steven Elwell (Senior Director of Digital Marketing), brought in Intero Digital to understand what wasn’t working.
The answer: 5 signals. Verification mechanisms that AI assistants apply, independent of PageRank.
Freshpet a découvert un paradoxe : top ranking n’égale pas visibilité AI. Voici comment l’équipe a diagnostiqué le problème et appliqué les 5 signaux en cascade pour retrouver sa place dans les AI Overviews.
Chronologie : du classement #1 invisible à la présence constante
Le parcours Freshpet depuis août 2026 : identification du problème et correction progressive
Freshpet, #1 organic… and completely invisible to AI
In August 2026, Freshpet shared its results at a Search Engine Journal webinar. The problem is clear: for queries like « best refrigerated dog food », their product page ranks first in organic.
Just above it, the AI Overview cites two or three competitors. Freshpet isn’t there.
This isn’t a malfunction. It’s an additional filter. Generative AI doesn’t reproduce the top 3 of the SERP. It selects brands it judges legitimate on the topic — broadly, not just on that specific page.
Intero Digital spotted a pattern. Brands appearing in AI Overviews systematically had:
- presence on reference sites (press, industry associations),
- consistent and complete structured data,
- unlinked mentions,
- verified reviews across multiple platforms,
- semantic coverage around the brand, not just around products.
Freshpet tackled these 5 axes in six weeks. Result: a citation rate multiplied by 4 on key queries (source: data shared by the Freshpet team on August 20, 2026).
Freshpet et nos clients ont vu des résultats spectaculaires en corrigeant ces 5 signaux. Voici l’impact mesurable : de l’invisibilité complète aux citations constantes en quelques semaines. Ces chiffres montrent pourquoi les anciens critères de ranking ne suffisent plus.
L’impact des 5 signaux de confiance sur les citations AI
Comparaison avant/après application des stratégies Freshpet et clients
Signal 1 — Backlinks are no longer a bonus, they’re a trust badge
We all lived through the PageRank era. Links were a vote. That’s still partly true, but today they’ve become an affiliation indicator.
AI doesn’t seek the best-linked page. It seeks the most-cited brand in serious editorial environments. Backlinks from sites like Petfood Industry, American Kennel Club, or university veterinary studies weigh far more than 300 links from average blogs.
Freshpet built a PR program strictly oriented toward expertise. No guest posts. No directory spam. Interviews, citations in « best of » articles, conference participations.
With my pet food client, I applied the same logic. We stopped producing generic content. We focused 3 months on getting 5 links from veterinary media. Result: 7 additional AI citations in 8 weeks.
Signal 2 — Structured data is no longer for the Knowledge Panel, it’s for the language model
Most product sheets I audited at Freshpet lacked Organization, a sameAs link to verified social profiles, consolidated aggregateRating.
LLMs exploit this markup to verify data consistency. Not for visual display. To cross-check and validate information.
Freshpet implemented complete JSON-LD:Organization → sameAs (verified LinkedIn, Facebook, Instagram), contactPoint customer service, brand on each product sheet. Everything was tested with Google’s validation tool and Rich Results Test.
This isn’t a detail. It’s the digital link that tells AI the page truly belongs to the brand it’s analyzing. One markup error, and trust collapses. I saw a site disappear from all AI citations for 3 weeks because a sameAs pointed to a deleted old Twitter account. The AI read that inconsistency as a bad signal.
Signal 3 — Unlinked mentions are the new popularity votes
This is the most underestimated signal.
When 15 articles mention « Freshpet » without adding a link, AI doesn’t ignore them. It counts them. It weighs them. A mention in Forbes is worth more than a link from a small blog.
Intero Digital mapped every Freshpet mention across the web: forums, press articles, Reddit posts, YouTube comments. They refocused efforts to multiply brand citations in non-promotional contexts.
With my client, I set up a basic Google Alerts and answered 12 technical questions on dog forums mentioning their brand naturally. After 6 weeks, 5 new AI citations without adding a single link.
Signal 4 — Verified customer reviews feed overall authority, not just conversion
An AI Overview doesn’t read your reviews page. It watches the constellation of reviews around your brand: Trustpilot, Google Merchant Center, Amazon, veterinary sites.
Freshpet launched a multi-platform review collection campaign, using solutions like Bazaarvoice and integrating aggregateRating feeds with identical markup across all channels.
Star count matters. But review freshness matters too. I observed that a brand with 430 reviews where 80% are over 12 months old gets twice fewer citations than a brand with only 120 reviews all posted in the last 3 months. Recency activates trust in language models.
Signal 5 — Thematic authority is not a buzzword, it’s an algorithmic measure
A site with 40 pages of siloed content on pet nutrition reads as a source. Not as a seller.
At Freshpet, the team created educational content at scale: articles on canine nutrition, food transition guides, sourced comparisons. All linked to product pages in a silo structure.
What I call thematic authority comes down to a ratio: the number of informative pages around a topic divided by the number of transactional pages. A ratio below 1.5 is an alarm signal for AI.
Freshpet raised this ratio from 0.8 to 2.7 in three months. Citations followed. With another pet health client, the same effort moved citations from 2 to 15 in 6 weeks.
This isn’t content for the sake of it. It’s proof of expertise in the model’s eyes.
Avant de lancer vos actions, testez-vous sur ces 5 dimensions. Cette grille interactive vous permet d’identifier où vous êtes faible et par où commencer. Un score inférieur à 3/5 sur une dimension = action prioritaire.
Auto-audit : vérifiez votre conformité AI
Évaluez votre marque sur les 5 signaux de confiance qu’exigent les AI Overviews
How to audit your AI visibility tomorrow
You don't need special tools. Here's the checklist I apply in my DOSE audits (framework taught by Guillaume Attias, BMO Academy):
- Search your brand + 10 main queries with the site:google.com/search?udm=50 option to force AI Overview display. Note whether your brand appears.
- Map your backlinks from the last 6 months. Segment sites with high thematic authority. Fewer than 3 links from this type? It's an action item.
- Validate your JSON-LD markup via Google's structured data test. Check for the absence of critical errors, and especially confirm a sameAs for each social profile.
- Set up your unlinked mention tracking with a Google Alert on your brand, and monthly monitoring. Each unlinked mention is an untapped lever.
- Assess your ratio of informative pages to product pages. Below 1.5, plan 5 to 7 expert content pieces in the next 45 days.
Freshpet executed these steps with rigor. I've reproduced this framework across 7 audits in September 2026. Results come fast when signals align.
DOSE audit of your AI citations in 45 minutes
I review your 5 trust signals live. Results sent within 24 hours. You'll see exactly why your competitors are cited instead of you.
Book a strategic call — 45 minFrequently Asked Questions
Is traditional organic ranking still useful for AI Overviews?
Yes, content needs to be relevant, but that's not enough to be cited. AI adds a trust filter. Without it, even ranking first changes nothing.
How long does it take to see results on AI citations after correcting signals?
<p>First results arrive between 3 and 8 weeks, depending on indexing speed for new mentions and Knowledge Graph update frequency. I've seen breakthroughs in 10 days.</p>
Do social media mentions count?
Yes, especially discussions on Reddit, YouTube, and TikTok when indexed. AI uses them as social proof, especially if the account sharing them is old and actively posting regularly.
Do AI Overviews replace traditional SEO or are they complementary?
Complementary. Well-ranked pages feed AI crawls, but only verified brands surface. SEO work remains the foundation.
Do you need a big budget to implement these 5 signals?
No. Structured data corrections cost nothing. Mention monitoring and refactoring your informative-to-product page ratio you manage in-house. The money goes mainly to PR with real experts, which you can start with a specialized freelancer.
