Blackout Research · Utah · 2026
Utah Google-to-AI Search Visibility Study 2026
A planned observational study of whether businesses visible in Google's organic results also appear when buyers ask major AI answer engines for recommendations.
Research Question
Does Google visibility carry into AI recommendations?
A business can be visible in traditional search without appearing in an AI-generated shortlist. Study #3 is designed to measure that gap using a fixed Utah sample, matched buyer-intent prompts, multiple AI answer engines, and repeated runs.
The study will compare Google's organic results with recommendations from ChatGPT, Gemini, Claude, and Perplexity. It will also record which sources the AI systems cite and how consistently the same businesses reappear when an identical prompt is repeated.
Planned Research Design
A controlled Google-to-AI comparison
These are study-design targets, not findings.
Planned Measures
What the study will measure
Search-to-AI overlap
How often a Google-visible sampled business is also named by an AI engine for the matched buyer intent.
Cross-engine consensus
Whether a business appears on zero, one, two, three, or all four tested AI engines.
Repeatability
Whether the same business appears in 0/3, 1/3, 2/3, or 3/3 repeated runs on the same engine.
AI-only discovery
How often AI recommends an eligible Utah business that is not present in Google's top ten organic results for the matched query.
Citation provenance
Whether visible sources come from the recommended business's own website, directories/review platforms, publishers, social/community sites, institutions, or other sources.
Official-site citation
How often a recommended business's own domain is visibly cited in the answer.
Research Strata
Six categories and six Utah search markets
Business categories
- HVAC / heating and cooling
- Personal injury law
- Dental practices
- Accounting firms
- Property management
- Managed IT services
The categories are research strata. The study will not rank categories as better or worse.
Utah search markets
- Salt Lake City
- Utah County
- Ogden
- St. George
- Logan
- Utah statewide
Sampling/search geography is a study variable and should not be interpreted as a claim about a business's complete customer territory.
Google Baseline
The sample starts with businesses already visible in Google.
For each frozen buyer-intent query, the study will capture the first ten organic Google results using a standardized, permitted collection method.
Directories, editorial lists, associations, government pages, duplicate domains, and businesses that do not meet the study's Utah operating-identity rules will be excluded from the candidate business frame.
A deterministic selection process will lock the target 120-business sample before AI collection begins.
AI Collection
The same buyer intent, repeated across AI answer engines.
Every frozen prompt will be tested in a clean conversation with live web search enabled where the platform provides it.
Core surfaces
- ChatGPT Search
- Gemini
- Claude with web search
- Perplexity Search
Run controls
Each prompt will be run three times on each engine. The study will preserve model/surface labels, run timestamps, visible recommendations, mention position, and citations.
Google AI Mode may be reported as a separate secondary surface only if it can be collected consistently across the full prompt set.
Repeated Runs
One answer is not a stable ranking.
AI answers are probabilistic and can change from one run to the next. A business that appears once may disappear on a repeat.
Instead of treating one generated list as a conventional search ranking, Study #3 will report repeatability: whether the same business appears in zero, one, two, or all three independent runs.
Citation Classification
A recommendation and a citation are different observations.
Visible sources will be classified as:
- Official business website
- Directory / review platform
- Editorial / news / publisher
- Social / community
- Government / institution / association
- Search / platform-owned property
- Other
- Unresolved
The study will not assume that a citation caused a recommendation merely because both appear in the same answer.
Research Safeguards
Public observation without synthetic customer activity
- No sampled company names in formal buyer-intent prompts
- No synthetic inquiries, calls, chats, bookings, purchases, or logins
- No paid-ad clicking
- No public best/worst business scorecard
- No unavailable output coded as a negative result
- No claim that visibility proves business quality
- No causal ranking claims from observational data
- No silent reruns until a preferred answer appears
- No invented model-version labels
- No Study #3 findings published before data and reliability gates pass
Reliability Plan
A delayed second coding pass
After first-pass coding is frozen, a deterministic subset of 6 prompt IDs will be selected for delayed blinded re-coding.
The second coding pass cannot begin for at least seven full calendar days.
Manual fields such as entity reconciliation, recommendation classification, mention position, citation-source type, and official-domain attribution will be checked for exact agreement.
Final findings will not publish unresolved manual-coding failures as though they were reliable measurements.
Planned Final Report
The same canonical page will hold validated findings.
After the formal collection and reliability gates pass, this same canonical page is expected to be updated with:
- validated executive summary
- Google-to-AI overlap
- engine-by-engine visibility
- repeatability distributions
- cross-engine consensus
- AI-only recommendations
- citation-source mix
- official-site citation share
- vertical and market observations
- reliability results
- exact denominators
- limitations
- citation guidance
- researcher/publisher attribution
No final findings should be inferred from the methodology release.
Related Blackout Research
The Blackout Utah research program
Related Blackout Expertise
Research context, not a sales detour
Citation and Research Use
Preferred citation for methodology release
Nieporte, Benjamin. “Utah Google-to-AI Search Visibility Study 2026: Methodology Release.” Blackout Marketing & Consulting, 2026.
Canonical: https://blackoutconsulting.com/pages/utah-google-to-ai-search-visibility-study-2026
Publication status: Research in Progress / Methodology Release. Formal findings have not been released.
Researcher and Publisher
Benjamin Nieporte · Blackout Marketing & Consulting
Benjamin Nieporte, Founder & Principal Strategist at Blackout Marketing & Consulting, is the principal researcher.
Blackout Marketing & Consulting is the publisher. Blackout is a founder-led growth marketing consultancy based in Lehi, Utah.
Inclusion in the research does not imply a client, partner, sponsor, or endorsement relationship.