AI Market Discovery Research
Overview
CiteWorks Studio's AI Market Discovery research tracks how brands appear across commercially relevant AI-assisted discovery, recommendation, pricing, and comparison experiences.
Each industry report uses a single evergreen URL that is updated as new benchmark data becomes available. Rather than publishing a new page every month, the report accumulates historical context, preserving important changes in recommendation coverage, rank, sentiment, competitive position, buyer intent, and source evidence over time.
This creates a research library built around unique category evidence, supported by shared methodology, metric, and research-standard resources.
How the Research Library Is Structured
Evergreen Industry Reports
Each industry/category report contains the evidence unique to that market, including:
- current recommendation coverage;
- historical movement from the original baseline;
- important monthly changes;
- competitive ranking changes;
- top-three and rank-one recommendation performance;
- sentiment movement;
- buyer-intent analysis;
- brand opportunity diagnostics;
- representative prompt evidence where available; and
- source/citation patterns where observable.
The same URL is updated over time so historical authority and longitudinal context accumulate on one page.
Shared Methodology
The AI Market Discovery Methodology explains:
- how the query universe is selected;
- what constitutes a prompt-surface observation;
- how brand relevance is determined;
- how buyer intent is classified;
- how recommendation outcomes are coded;
- how the qualified benchmark denominator is created; and
- how QA and longitudinal updates are managed.
Shared Metric Definitions
The AI Visibility Metric Definitions provide stable definitions and interpretation rules for:
- valid recommendation coverage;
- top-three recommendation rate;
- rank-one recommendation rate;
- raw mention presence;
- net sentiment;
- qualified surface breadth;
- valid recommendation shortlist share; and
- modeled AI Authority Value.
Shared Research Standards
Research Standards & Interpretation defines the universal boundaries and publication controls that apply across the research program, including:
- what the benchmark can and cannot establish;
- dynamic AI-output limitations;
- small-count interpretation;
- longitudinal integrity;
- methodology versioning;
- source/citation interpretation;
- evergreen report policy; and
- publication QA standards.
How to Read an AI Market Discovery Report
A category report is intended to answer progressively deeper questions.
1. Who is visible?
Recommendation coverage and presence metrics show which tracked brands are entering commercially relevant AI answers.
2. Who is preferred?
Top-three and rank-one rates distinguish brands that are merely included from brands that AI systems place prominently or recommend first.
3. What changed?
The evergreen report compares the current month with the original baseline and prior history, highlighting material gains, declines, leadership changes, and reversals.
4. Where is the movement happening?
Buyer-intent, prompt, and surface analysis can show whether a brand performs differently in broad discovery, pricing/value evaluation, or direct comparison.
5. What evidence surrounds the recommendation?
Where citations or attributable sources are available, source analysis can identify the external information environment associated with AI recommendations.
6. What should a brand investigate next?
The public benchmark identifies where the competitive signal is strongest. Company-level analysis can then determine the prompts, competitors, surfaces, attributes, and source gaps responsible for that position.
Research Scope: Raw Collection vs. Qualified Benchmark
AI Market Discovery reports do not treat every generated AI answer as equally useful for competitive analysis.
A typical monthly category run begins with approximately 800 prompt-surface observations generated from high-demand category queries across the benchmark's defined AI/search surface universe.
Those observations are then narrowed through two stages:
- Tracked-brand relevance — does the answer meaningfully surface one of the tracked competitors under the benchmark rules?
- Commercial buyer intent — does the observation fit Brand Recommendation, Pricing & Value, or Multi-Brand Comparison intent?
Only observations that satisfy the applicable qualification rules enter the public benchmark denominator.
This approach is designed to focus the reports on AI interactions where a consumer is plausibly discovering, evaluating, or choosing among brands.
Reporting Changes Clearly
When a percentage rate changes, reports describe the difference in percentage points.
For example, if recommendation coverage moves from 27.3% to 44.8%, the report describes that as an increase of 17.5 percentage points.
For compact tables and charts, this appears simply as Up 17.5 points.
For Brands and Marketing Teams
The public research is designed to show where a brand sits in the AI-mediated competitive landscape.
For a company appearing in one of these reports, the next questions are typically more specific:
- Which high-intent prompts produce our strongest visibility?
- Where are competitors replacing us in the recommendation order?
- Which buyer-intent clusters are we winning or losing?
- Which AI/search surfaces consistently include us?
- Which product attributes does AI associate with our brand?
- Which third-party sources appear around recommendations in our category?
- Are competitors supported by stronger evidence in high-demand prompt families?
A company-specific AI visibility audit uses the underlying prompt, surface, competitor, ranking, sentiment, and source data to investigate those questions.
Request an AI visibility audit →