AI Visibility Audit · Specialty CPG Series

Functional & Premium
Protein Bar Audit

Which brands AI recommends — and the training data gap no one is talking about

PublishedMay 2026
PlatformsClaude · ChatGPT · Gemini · Perplexity
Prompts Run50 prompts · 5 clusters · 2× averaged
MarketUnited States

Key Findings

Methodology: Queries were run via API across Claude, ChatGPT, Gemini, and Perplexity — not consumer web interfaces. Prompts structured across five purchase-intent clusters: Health & Better-For-You, Flavor & Quality, Clean Ingredient & Performance, Retail & DTC Discovery, and Subscription & Repeat Purchase. Each prompt run twice and results averaged. Entity normalization: Raw NER produced over 600 unique brand strings. Variants such as "RXBAR," "RXBARs," "RX Bar," and "RXBAR Collagen" consolidated to a single canonical entity. Retailers, subscription box services, and media publications filtered.

Platform Divergence —
Top 20 Brands

Gemini leads overall platform volume at 30.9%, followed by ChatGPT (26.3%), Perplexity (22.1%), and Claude (20.6%) — the first audit in this research series where Claude ranks last by platform volume. Platform divergence patterns in this category are among the most varied of any CPG audit conducted to date.

BrandClaudeChatGPTGeminiPerplexityTotal
Quest2352763721731,056
RXBAR2782952322201,025
ONE Bar141199349224913
ALOHA111162355274902
KIND213375196106890
GoMacro15327027065758
Built Bar9711425577543
Clif792188789473
Barebells9226194153465
Larabar2071731623419
Bulletproof986510583351
No Cow971456246350
Primal Kitchen278815527297
EPIC361714917273
Grenade203418015249
Perfect Bar65457549234
Orgain69568920234
Vital Proteins38476720172
Pure Protein3267104140
Garden of Life2039584121
Larabar scores 207 on Claude and 173 on ChatGPT but only 16 on Gemini — an inverse of the expected SEO-lift pattern for a legacy brand with strong retail distribution. Barebells scores 194 on Gemini but only 26 on ChatGPT. Pure Protein is almost entirely Perplexity-driven (104 of 140 total). GoMacro and Built Bar have strong trained-platform visibility but significant Perplexity gaps.

Meaningful visibility elsewhere —
zero on one platform.

These gaps represent the most structurally addressable visibility risks in the category. David Bar is a special case — not a content gap but a temporal one.

BrandMentions ElsewhereGap PlatformGap Mentions
Larabar396Gemini16
Misfits90Claude0
Oatmega75Gemini0
Legion59ChatGPT0
David Bar57Claude · ChatGPT · Gemini0
Thunderbird48Gemini0
Ancient Nutrition47Perplexity0
Transparent Labs47Claude · ChatGPT · Gemini0
IQBAR45Claude0
TRUBAR33Claude0
Further Food26ChatGPT · Perplexity0
NuGo26ChatGPT0
David Bar is a training data gap story: 57 mentions entirely from Perplexity, zero on Claude, ChatGPT, and Gemini. Launched in 2024, it postdates the training cutoffs of the three knowledge-based platforms. It surfaces only through Perplexity's live web retrieval — visible in real time but absent from trained recommendation knowledge. For any brand launched in 2023 or later, this dynamic is the most urgent AI visibility issue to understand.

How the category
splits by intent.

Functional protein bar prompts do not return a single consistent brand set. Five purchase-intent clusters reveal meaningfully different competitive landscapes — with ALOHA the only brand threading across multiple cluster tops.

Cluster 01
Health & Better-For-You
The functional buyer: collagen, gut health, clean protein, everyday wellness
ALOHA Bulletproof KIND RXBAR GoMacro

ALOHA leads at 207 mentions — the only plant-based brand to top any cluster in this audit. Its position reflects consistent cross-platform presence on wellness and clean protein queries. Bulletproof's 204 mentions are driven almost entirely by its collagen positioning — a narrow but deeply reinforced content signal that AI systems surface reliably for collagen-specific queries.

Cluster 02
Flavor & Quality
The taste-first buyer: indulgent, satisfying, premium flavor experience
ONE Bar Quest Barebells ALOHA RXBAR

ONE Bar leads at 281 mentions — its highest cluster performance and a reflection of how consistently AI systems associate the brand with dessert-forward, indulgent protein formats. Quest follows at 273, driven by years of flavor innovation content absorbed by trained platforms. Barebells at 192 punches above its U.S. market share, suggesting its European brand equity translates well into AI training data. ALOHA's 186 makes it the only brand in the top five of both Health and Flavor clusters — a dual-positioning signal that is rare and commercially significant.

Cluster 03
Clean Ingredient & Performance
The macro buyer: athlete, keto, paleo, grass-fed, complete amino profiles
ALOHA RXBAR GoMacro EPIC ONE Bar

ALOHA and RXBAR lead — both reflecting brands whose ingredient transparency is a core content signal rather than a marketing claim. RXBAR's "No B.S." positioning — built around a short, readable ingredient list — is one of the most well-documented ingredient narratives in the category and AI systems surface it consistently on clean-label and performance queries. EPIC's strong showing in this cluster is driven by meat-based protein and paleo content that trained platforms recognize as distinct from the broader bar category.

Cluster 04
Retail & DTC Discovery
The active buyer: where to purchase, which stores carry it, DTC options
Quest RXBAR KIND ALOHA GoMacro

Retail and DTC queries favor the brands with the broadest physical distribution. What's notable is how closely this cluster mirrors the overall rankings — brands visible in health and flavor queries are also visible when buyers ask where to purchase. DTC-native brands without strong retail presence are largely absent here, suggesting that AI systems default to distribution breadth as a proxy for purchase accessibility.

Cluster 05 · Most Revealing
Subscription & Repeat Purchase
The loyalty buyer: monthly delivery, subscribe-and-save, curated bar boxes
Quest RXBAR ONE Bar KIND Built Bar

Quest leads with 326 mentions — its highest cluster performance — despite the fact that several smaller DTC-native brands offer more developed subscription models. RXBAR follows at 249, KIND at 195. None of these brands are known primarily for subscription innovation; all three are mass-market bars available in grocery and convenience retail. Meanwhile, brands with genuinely differentiated DTC subscription models surface far below their program quality would suggest. This gap reveals how AI systems answer subscription questions: by defaulting to the brands they know best, not the brands with the most relevant purchase infrastructure. The content architecture that would signal a superior subscription program to AI systems is almost universally underdeveloped across this category.

Three signal types account for
the majority of high-visibility patterns.

The brands underperforming on Claude and ChatGPT relative to their market position are not underperforming because their products are unknown. They are underperforming because the evidence for why their bar is worth buying is either absent, buried in architecture AI systems cannot reach, or present only in formats that trained platforms did not encounter during training.

Signal 01
Ingredient Narrative Depth

The brands AI recommends most consistently are the ones that turned their ingredient philosophy into a content system, not a label claim. RXBAR built its entire brand identity around a readable ingredient list and has produced years of content explaining what that means and why it matters. ALOHA has done the same for plant-based protein completeness and organic sourcing. GoMacro has built out macrobiotic and sustainability content that AI systems recognize as authoritative.

In each case, the ingredient story is not confined to a product page — it is distributed across cluster-relevant content that trained platforms encountered and weighted during training.

Brands with strong formulation credentials but thin editorial content — notable collagen, seed-based, and functional ingredient bars in this audit — surface at a fraction of the volume their product quality would warrant. The science exists in the formulation. It has not been written into content that AI systems can find and cite.
Signal 02
Cross-Cluster Presence

Single-cluster brands are invisible to buyers who don't already know what they want. Most brands in this audit surface consistently in one or two clusters — the clusters that most closely match their primary positioning. But purchase decisions are rarely that linear. A buyer who starts with "best collagen protein bar" may follow up with "which ones taste good" or "where can I buy them."

ALOHA is the clearest example in this dataset of a brand that accompanies a buyer across the decision journey rather than appearing once and disappearing.

Brands that have not produced content addressing adjacent intent types — the flavor story for a functional bar, the performance story for a taste-first brand — are leaving significant cross-cluster visibility on the table.
Signal 03
Platform-Specific Content Architecture

SEO gets you Gemini. Content architecture distributed across the web gets you the other three. Gemini leads platform volume at 30.9% in this audit — brands with strong SEO and well-structured sites receive a natural lift there. Claude and ChatGPT draw from trained knowledge. Perplexity retrieves live web results in real time. Each platform has a different answer to the question of what makes a brand visible.

David Bar's Perplexity-only presence is the most dramatic illustration of this dynamic — and the clearest signal that any brand launched in 2023 or later faces a training data challenge that cannot be solved by SEO alone.

Brands that have invested heavily in SEO without a parallel investment in content depth and distribution will continue to see Gemini overperformance and Claude/ChatGPT underperformance relative to their actual market position.

David Bar launched in 2024, has generated significant press, and has zero mentions on Claude, ChatGPT, or Gemini. It surfaces only through Perplexity's live web retrieval. For any brand launched in 2023 or later, this is the most urgent AI visibility issue to understand — and the one most brands in this audit have not yet addressed.

The AI recommendation playing field
is more level than the search ranking field.

Research from Princeton and IIT Delhi published at KDD 2024 found that generative engine optimization produces disproportionate benefits for lower-ranked sites — with sites ranked fifth in search seeing up to 115% improvement in AI-generated response inclusion after content restructuring. For emerging and mid-tier protein bar brands competing against Quest and RXBAR, this is the most actionable competitive intelligence available.

Three content types move the needle most in this audit. Ingredient rationale published at depth zero or one — on product pages, not blog posts — is the single highest-leverage change for brands with strong formulations and thin product copy. Founder and use-case narrative that connects personal motivation to the specific buyer experience produces the cross-cluster presence the top brands in this dataset demonstrate. And subscription and DTC-specific content that tells AI systems what the purchase experience looks like is almost entirely absent in this category and represents the largest untapped cluster opportunity.

The window to establish position is open now. The AI recommendation playing field is more level than the search ranking field — and the brands that understand that first will hold positions that others will spend years trying to close.

This report is part of an ongoing series examining AI recommendation patterns across premium food, beverage, and hospitality categories. Ally Kiel Consulting publishes original audit data to help founders and operators understand how AI systems currently classify and recommend their brands — and what drives the gaps.

View all research →