analyst-common

To prevent AI hallucinations during web research, you can enforce direct web-search calls, require verbatim quotations of numeric data, and mandate cross-source verification across at least three independent sources with a ±1% numeric agreement threshold.

Cross-validate numeric data by capturing the original_text verbatim, acquiring at least three independent sources with recorded URLs and dates, and checking that the numeric values agree within a strict ±1% threshold.

The best way to ensure source citation traceability is to enforce mandatory web-search tool invocation and record verbatim original_text alongside the exact source URLs and retrieval dates for every numeric data point cited.

Yes, cross-source verification can be applied across sector-analyst, risk-analyst, rate-analyst, and macro-critic workflows to ensure reliable web data and prevent misinterpretation during information synthesis.

AI analysts hallucinate when synthesizing market data because they lack enforced direct web-search usage and verbatim quotation requirements, leading to unverified data interpretation without traceable source provenance.

Relying on a single web source for numeric data extraction lacks cross-source validation, meaning you cannot confirm data consistency within a ±1% agreement threshold or ensure the accuracy required to prevent AI hallucinations.