AI Disclosure

How proofnest uses artificial intelligence

How reports are generated

proofnest reports are generated entirely by artificial intelligence. No human analyst reviews individual reports before delivery.

Our system works in multiple stages:

Your input (idea description, market, audience) is sent to our research pipeline
AI agents scan dozens of live data sources — search engines, social platforms, app stores, review sites, startup databases, and more
Collected data is analyzed, cross-referenced, and structured into report sections
A self-challenge layer reviews findings for contradictions and weak evidence
A weighted scoring algorithm produces the final GO / YELLOW / RED verdict
The completed report is delivered to your dashboard

AI models and data sources

proofnest uses commercial AI models and licensed data APIs to generate reports. We do not disclose specific model names or API providers to protect our methodology, but we can confirm:

Large language models are used for analysis, synthesis, and report writing
Search and trend data comes from licensed commercial APIs
Social media and review data is collected from publicly available sources
All data sources are accessed through authorized API agreements

What AI-generated means for you

AI-generated content may contain inaccuracies or outdated information. While our multi-layer verification process significantly reduces errors, it does not eliminate them entirely.

You should:

Verify critical findings independently before making business decisions
Treat the verdict as one input among many, not as a definitive answer
Check source links provided in the report to validate key claims
Use reports as a starting point for deeper research, not as a replacement for it

Source attribution

Every factual claim in a proofnest report is linked to its data source wherever possible. Source links appear inline throughout the report. If a source link is broken or unavailable, it means the original page may have been updated or removed since the report was generated.

Self-challenge layer

Our AI does not simply compile information. At each stage of analysis, the system actively looks for contradictions, challenges its own conclusions, and flags weak evidence. This self-challenge process is designed to reduce confirmation bias and produce more honest assessments.

This is not the same as human review. It is an algorithmic safeguard, not an editorial process.

Data handling

Your idea descriptions and report inputs are used solely to generate your report. We do not share your specific ideas with other users or make them publicly available.

After report generation, we extract anonymized, aggregated market data (competitor names, market sizes, demand signals) into an internal knowledge base. This data contains only publicly available information and is stripped of any connection to your specific idea. This practice is described in our Privacy Policy and Terms of Service.

Limitations

Market data has a shelf life. Reports reflect data available at the time of generation. For fast-moving markets, we recommend refreshing your research every few months.
Data depth varies by niche. Some markets have more publicly available information than others. Reports always work with what's available — and clearly note when data is limited.
AI analysis, not absolute truth. Our multi-layer verification reduces errors significantly, but no research tool is perfect. Source links are provided so you can verify the findings that matter most to you.
Verdicts are assessments, not predictions. GO, YELLOW, and RED reflect the strength of current evidence. Markets evolve, and timing matters — use the verdict as a strong signal, not a guarantee.

Feedback and corrections

If you find an error in a report or believe a section contains inaccurate information, use the "Report Issue" button available on every report page. We review all feedback and use it to improve our system.

Questions

For questions about how proofnest uses AI, contact us at [email protected].