Governance
AI Usage Policy
How Belvedere uses machine learning while keeping evidence and human review at the center of research.
Belvedere Research uses technology and machine-learning tools primarily as research infrastructure: to assist with numerical extraction, data normalization, calculations, filing monitoring and quality-control workflows. Our investment analysis, research judgments, conclusions and final editorial decisions remain the responsibility of our human research team.
AI systems can produce errors, including extraction and interpretation errors. Accordingly, machine-generated outputs are treated as inputs to a controlled research workflow, not as authoritative sources. Financial facts are reconciled to primary documents before publication.
Where AI assists with drafting, it does so inside that workflow. AI does not independently author or determine the investment thesis. No machine-generated report is published without human review.
See methodology for calculation and source rules.
Frequently asked questions
- How does Belvedere use AI?
- For infrastructure: extraction, normalization, calculations, monitoring and quality-control assistance. Not as the source of financial facts or as the decision-maker on research conclusions.