The fine print.

This disclosure applies to all outputs generated by the BrandCapsule AI Discovery Intelligence platform, including the AI Discovery Snapshot, Brief, Lab, Actions, and AI Discovery Analyst briefings (collectively, "platform outputs").

BrandCapsule is committed to transparency about how our platform generates outputs and the inherent limitations of AI-generated content. Please read this disclosure before using or distributing any platform output.

1. How BrandCapsule Generates Outputs

BrandCapsule queries the following third-party AI language model APIs to generate platform outputs:

• ChatGPT — operated by OpenAI, Inc.
• Claude — operated by Anthropic, PBC
• Gemini — operated by Google LLC

BrandCapsule applies its patent-pending methodology to structure, score, and validate the raw outputs returned by these models. Our methodology operates in six discrete, individually constrained stages — Extraction, Structural Analysis, Constrained Scoring, Cross-Model Validation, Aggregation, and Evidence Layer — designed to reduce the variance inherent in probabilistic AI outputs and produce reproducible, evidence-backed results.

The characterizations of brands, competitors, market positions, and buyer behavior contained in platform outputs are AI-generated. They are not independently verified market research, consumer surveys, or statements of fact.

2. What "Verified, Not Asserted" Means

Our verification is evidence-based. Before we render a score, we check that the AI actually named your brand (if fewer than three AI responses mention you, we render ‘UNKNOWN’ rather than guess a score); that certain reputation signals include cited evidence in the AI’s own answer (or we render ‘UNKNOWN’); and that the AI grounded its response in live web sources we can capture. When we detect a contradiction between an AI’s claim and information you’ve submitted about your brand, we surface it in the report’s fact-verification section — we do not silently reject the AI’s claim on your behalf, because part of the value of the report is showing you what AI models actually say about your brand, even when they’re wrong.

This means:

• Claims about your brand that contradict your own submitted data are flagged in the report’s fact-verification section — they are surfaced to you, not silently removed.
• Claims about third-party brands, competitors, and vendors are not independently verified against those companies’ own data. Third-party characterizations in your reports reflect what AI models generate — not confirmed facts about those companies.
• The verification step reduces but does not eliminate the risk of inaccurate outputs. AI models can generate internally consistent but factually incorrect content that passes structural validation.

3. Known Limitations of AI-Generated Content

• Hallucination. AI models can generate plausible-sounding but factually incorrect information, scores, rankings, or associations between companies that have no basis in reality.
• Misattribution. AI models may assign characteristics, reviews, incidents, or competitive attributes of one company to another with a similar name or in a similar category.
• Training data cutoffs. AI outputs reflect training data as of each model’s knowledge cutoff date, which may be months or years in the past.
• Model variance. The same query submitted to the same model may return materially different outputs on different days or across different model versions. This is inherent to probabilistic language models.
• Cross-model disagreement. ChatGPT, Claude, and Gemini may produce substantially different characterizations of the same brand for the same buyer question. BrandCapsule surfaces this variance; it does not resolve which model is correct.
• Representation bias. AI models reflect biases in their training data, which may include over- or under-representation of certain industries, geographies, brand sizes, or time periods.

4. Output Types

A one-time baseline diagnostic reflecting a single point-in-time query across ChatGPT, Claude, and Gemini. Subject to all limitations described above. The Snapshot is intended as a starting point for understanding your brand’s AI visibility, not as a definitive market study.

A weekly narrative summarizing your AI Discovery position, recommendation rate, and discovery map. Reflects the most recent query cycle. AI model behavior may have changed since prior weeks’ results.

Deep competitive and source analysis identifying buyer questions where competitors are recommended and surfacing the sources AI trusts. Source citations in Lab outputs reflect what AI models attribute as sources, which may not match actual current source content or accuracy.

Prioritized improvement programs including content wins, campaigns, and GEO site audits. Recommendations are derived from AI model outputs and BrandCapsule’s analytical framework. They represent directional intelligence, not guaranteed outcomes. Publishing content in response to Actions recommendations is subject to AI model reindexing timelines outside BrandCapsule’s control.

The AI Discovery Analyst is a human-assisted advisory layer. Analyst briefings reflect a human advisor’s interpretation of AI-generated platform data. Analyst recommendations are advisory in nature and do not constitute legal, financial, investment, or professional marketing advice.

5. Appropriate Uses

• Directional intelligence for internal brand and marketing strategy decisions.
• Identifying buyer questions and competitive gaps to prioritize content and positioning work.
• Tracking trends in AI perception of your brand over time across multiple query runs.
• Internal presentations and board updates with appropriate disclosure of AI-generated nature.
• Client deliverables prepared by agencies with appropriate attribution and AI-generated content disclosure.

6. Inappropriate Uses

• Publishing AI-generated scores or characterizations of third-party brands as independently verified facts in external marketing materials, press releases, or public statements.
• Using AI-generated competitive characterizations as the basis for legal, regulatory, or compliance proceedings against third parties without independent verification.
• Making advertising claims about competitors based on platform outputs without independent substantiation.
• Presenting AI Discovery Index scores or AI recommendation rates as the result of consumer surveys, independent audits, or human market research without disclosure.

7. Third-Party Concerns about AI-Generated Characterizations

BrandCapsule does not create, author, or endorse the characterizations that appear in our reports. Our platform aggregates and analyzes what third-party AI systems (ChatGPT, Claude, Gemini) say about companies and brands — these characterizations are generated by those AI models, not by BrandCapsule, and reflect those models’ own outputs, which we do not control and cannot alter.

If you are a company, brand, or individual who disputes an AI-generated characterization appearing in a BrandCapsule report, you may submit your concern to us. Submitting a concern does not obligate BrandCapsule to remove, alter, or correct the underlying characterization, as we have no ability to change what a third-party AI model generates.

If you have questions about any of these policies, please contact us at:

hello@brandcapsule.ai
BrandCapsule, Inc.
343 Soquel Avenue, Suite 305
Santa Cruz, CA 95062