AI RECOMMENDATION INDEX
How AI assistants recommend businesses.
AI See You runs structured buyer questions across major AI assistants every week to understand which businesses get recommended and which signals influence those decisions.
AI See You operates a structured research program studying AI recommendation behaviour across industries, geographies, and platform types.
To date, the AI See You research program has analysed 50,000+ real buyer queries across ChatGPT, Claude, Gemini, and Perplexity across two structured tracking runs. This figure updates as new runs complete.
The gap between search presence and AI recommendation.
A business can rank well in Google search and still have a significant Recommendation Gap. The Recommendation Gap is the distance between how well-known a business is to human searchers and how confidently AI assistants recommend it. Most businesses have invested years in search optimisation. Almost none have invested in recommendation infrastructure. As AI assistants become the primary discovery layer for buyers, the Recommendation Gap becomes the most important metric a business can track. AI See You measures and closes the Recommendation Gap.
DEFINITION
The Recommendation Gap is the measurable distance between a business's current AI recommendation rate and its potential recommendation rate once recommendation infrastructure is fully deployed.
RECOMMENDATION FREQUENCY
Which businesses appear in AI answers
We track how often businesses in each industry category appear when real buyer questions are asked across major AI assistants. Patterns emerge quickly about which signal types correlate with consistent recommendation.
SIGNAL INFLUENCE
Which signals AI assistants rely on
Not all signals carry equal weight. Our research identifies which types of structured content, authority markers, and proof elements appear most consistently in businesses that receive strong AI recommendations.
CATEGORY COVERAGE
Where AI recommendations are already dominant
AI recommendations are not evenly distributed across industries. Some categories show high AI recommendation activity already. Others are largely uninfluenced. We map where the Recommendation Gap is widest.
PLATFORM VARIATION
How recommendations differ across AI assistants
ChatGPT, Claude, Gemini, and Perplexity do not always recommend the same businesses. We track where platform behaviour diverges and what that means for recommendation infrastructure.
Early patterns from the research program.
Initial findings from the first tracking runs. All patterns are indicative and will be updated as the dataset expands. This is a living document.
The findings below are based on initial structured tracking runs across a small cohort of businesses. They should be treated as directional indicators, not statistical conclusions.
PROFESSIONAL SERVICES
ACCOUNTING | AU and UK markets
Initial analysis suggests AI assistants frequently recommend accounting firms with structured service specificity. Firms describing themselves as generalist practitioners appear less consistently than firms with clear industry specialisations such as construction accounting, medical practice accounting, or ecommerce accounting.
INITIAL - EXPANDINGPROFESSIONAL SERVICES
FINANCIAL PLANNING | AU market
Financial planning businesses with clearly structured service descriptions, credential documentation, and client outcome framing appear more consistently in AI recommendation responses than those with generic positioning. Regulatory clarity such as AFSL details and service scope appears to correlate with recommendation frequency.
INITIAL - EXPANDINGHEALTH PRACTITIONERS
PHYSIOTHERAPY | AU and UK markets
Physiotherapy practices with structured local signals, visible practitioner credentials, and condition-specific content appear more frequently in AI recommendation responses. Practices describing only general physiotherapy services show a larger Recommendation Gap than those with structured condition and treatment pages.
INITIAL - EXPANDINGRETAIL AND CONSUMER
DTC BRANDS | AU market, initial
Consumer brands with comparison content, honest product descriptions including limitations, and use-case specific pages appear to be cited more frequently in AI product recommendation responses. Brands with only standard product listing content show a significant Recommendation Gap against category competitors with structured comparison and buying guide content.
INITIAL - EXPANDINGHow the research program works.
Structured question sets
We define a set of buyer questions for each industry category reflecting the real prompts buyers use when asking AI assistants for recommendations.
Multi-platform tracking
Each question set runs across ChatGPT, Claude, Gemini, and Perplexity. We record which businesses appear, how they are described, and what signals appear to influence the recommendation.
Iterative refinement
Findings are treated as hypotheses, not conclusions. As the research network expands, patterns are tested, refined, and updated. The index is a living document.
Receive index updates when new findings publish.
New industry findings, platform behaviour changes, and quarterly summary reports are sent to the list first.
Frequently Asked Questions
The AI Recommendation Index is a continuously updated research publication from AI See You tracking how AI assistants recommend businesses across industries. It documents patterns in recommendation frequency, signal influence, category coverage, and platform variation. It is updated as new tracking data is published.
The Recommendation Gap is the measurable distance between a business's current AI recommendation rate and its potential recommendation rate with full recommendation infrastructure in place. It is measured through the Recommendation Score, which runs structured buyer queries through major AI assistants and records whether the business is recommended.
Current findings are based on initial tracking runs across a small research cohort. They are directional indicators, not statistical conclusions. All findings are clearly labelled. As the research network expands, findings will be updated and refined. The index is a living document.