Case Study

From 36 to 58/100: How We Improved nutriish.me's AI Visibility — and What We Learned

Written by Joelle Berger · Strategic Pilot at AI Audit Scan
J

Joelle

2026-02-23 · 6 min read

Nutriish, Your Private Healthy Eating Concierge

nutriish.me — Digital Nutrition Concierge for families and professionals

When we first scanned Nutriish, the score was 36/100. Not because the product is bad — it is a well-designed online nutrition concierge for families and professionals. The problem was purely technical: the site was structurally invisible to AI systems like ChatGPT, Claude, and Gemini.

This case study documents two phases of optimization, a live citation tracking run across 4 AI engines, and one honest failure. Every step is tracked live at aiauditscan.com/case-study/nutriish.

Visit nutriish.me

Online Nutrition Concierge for families and professionals

The Baseline: 36/100

The initial scan on February 22, 2026 revealed a site that was technically functional for human visitors but structurally invisible to LLMs. The homepage had 47 words, far below the minimum threshold for AI content extraction. There was no structured data of any kind, no robots.txt, and no canonical tag.

SignalStatusImpact
JSON-LD structured dataMissingLLMs cannot identify entities
Schema markupMissingNo Organization, Service, or FAQ context
Word count47 wordsBelow LLM extraction threshold
robots.txtMissingAI crawlers may skip site
Canonical tagMissingDuplicate content risk
H2/H3 headings0No content structure for extraction
Key insight: A score of 36/100 does not mean the website is broken. It means AI systems have no reliable way to understand what the site is, who it serves, or what it offers. The site exists on the web but not in AI memory.

Phase 1: Technical Foundation (February 2026)

In a single session on February 23, 2026, we deployed five structural changes that moved the score from 36 to 57/100 — a 21-point gain without touching the visual design.

SignalBeforeAfter
JSON-LD structured dataMissingOrganization + WebSite + Service
FAQ SchemaMissing4 structured Q&A pairs
Word count47 words342 words
robots.txtMissingLive with AI crawler permissions
Canonical tagMissingAdded
H2 headings05 structured headings

Phase 2: Authority and Content Signals (February — March 2026)

Following the initial session, we implemented a second wave of improvements targeting Publisher Signals and Entity Clarity — the two lowest-scoring pillars in the AEO breakdown.

ActionPillar targetedStatus
Article schema added to blog postsPublisher SignalsDeployed
Testimonials with Review schemaE-E-A-T / TrustDeployed
Automated RSS news fetcherContent freshnessLive (Tue/Fri cron)
Blog post with case study contentPublisher SignalsPublished
Product schema on service pagesStructured DataDeployed
Wikidata entity entryEntity ClarityDeleted by Wikidata moderators
LinkedIn company pagePublisher SignalsPending
Honest failure — Wikidata deletion: We created a Wikidata entry for Nutriish to establish machine-readable entity disambiguation. It was deleted by Wikidata moderators citing insufficient third-party coverage. This is a known constraint for early-stage brands: Wikidata requires verifiable references from independent publications before an entry can be maintained. The fix is PR-first — earning coverage in nutrition and wellness media before attempting re-entry.

The March 2026 Score: 58/100

The March 18, 2026 scan returned a global AEO Index of 58/100 — above the Nutrition & Wellness industry average of 32/100, but still 7 points below the top-quartile threshold of 65/100. The two critical gaps remaining are Publisher Signals (0/10) and Content Structure (3/25).

4-Engine Citation Tracking: 85% Overall

On March 18, 2026, we ran a full AI Citation Tracking run across 4 engines — 20 queries total, 5 per engine — using queries calibrated for a digital nutrition service (not a physical wellness clinic).

EngineCitation rateQueries citedKey finding
ChatGPT100%5/5Cites Nutriish consistently vs MyFitnessPal
Claude100%5/5Cites but flags limited third-party verification
Gemini80%4/5Strong on comparison queries, one entity gap
Perplexity60%3/5Confuses Nutriish with Nourish — entity gap
Perplexity diagnostic: Perplexity's RAG engine returned: 'I believe you are asking about Nourish (not Nutriish), an online dietitian platform.' This is the entity disambiguation problem in action. Without a Wikidata entry or coverage in health publications, Perplexity cannot distinguish Nutriish from phonetically similar brands. This is the primary blocker for the remaining 40% gap on Perplexity.

What Comes Next

The path to 65/100 and full Perplexity citation runs through three actions: earning coverage in at least two independent nutrition or wellness publications (prerequisite for Wikidata re-entry), publishing long-form content with question-based H2 headings to address the Content Structure gap, and activating a LinkedIn company page to add the final Publisher Signal.

The citation data confirms that ChatGPT and Claude are already reliably citing Nutriish on digital nutrition queries. The remaining opportunity is Perplexity — which requires off-site authority, not more on-site technical fixes.

Live tracking: Score and citation data update automatically after each scan. View the full history at aiauditscan.com/case-study/nutriish.

What This Means for Your Site

nutriish.me is not an outlier. Based on our analysis of 5,609+ websites across 15 industries, the average AI readiness score is 38/100. Most sites that perform well in traditional search are still invisible to AI systems because they were never built for LLM extraction.

The fixes are not radical redesigns. They are structured additions that sit alongside your existing content without changing what your visitors see. A properly structured site can move from invisible to citable in days. The harder work — earning the off-site authority that source-based engines like Perplexity require — takes longer, but it starts with knowing exactly where you stand.

If you want to know where your site stands, run a free audit at aiauditscan.com. It takes 30 seconds and gives you a score with specific, actionable recommendations.

Ready to Check Your AI Visibility?

See how your e-commerce site compares to the 409 sites I analyzed. Get a detailed AI readiness report in seconds.