2026 Credit Union & Fintech AI Search Visibility Study: Measuring Total Brand Omission, Cross-Engine Disagreement, and Loan-Tier Drift Across ChatGPT, Gemini, Perplexity, and Claude (Zenodo (CERN European Organization for Nuclear Research))

Summary

• Audited institutions suffered an aggregate 96.0% Total Brand Omission rate across 960 brand-inclusion evaluations on major AI search platforms.p.1
• The Credit Union cohort recorded a 95.0% Total Brand Omission rate, while the Ethical Fintech cohort suffered a 97.1% Total Brand Omission rate.p.1
• Across 336 recommendation slots evaluated across 20 consumer queries, the study found a 95.7% Cross-Engine Disagreement rate.p.2
• Third-party lead-generation aggregators, comparison middlemen, and government program acronyms captured 17.3% of all AI recommendation slots.p.2
• A parameter sensitivity test revealed that 85.7% of recommended lenders vanished or rotated when adjusting auto loan amounts in $10,000 increments.p.2

Key insights

1. Audited institutions captured just 38 mentions out of 960 opportunities, yielding a 4.0% total AI search visibility rate.p.1
2. All four AI platforms agreed on the exact same recommended entity only 4.3% of the time across 234 unique query-entity combinations.p.2
3. The top four U.S. megabanks combined captured only 5.4% of total AI recommendation slots on transactional consumer queries.p.2
4. Three out of twelve audited institutions suffered a complete 100.0% Total Brand Omission rate with zero mentions across all 80 evaluations.p.9
5. Anthropic Claude returned zero lender recommendations on credit-tiered auto loan queries due to strict liability guardrails suppressing unverified rates.p.16
6. Only 1 out of 7 recommended lenders maintained consistent visibility across $60,000, $50,000, and $40,000 auto loan prompt tiers.p.18