The Verification Workflow

Verification should not be a separate phase that happens after research — it should be embedded in the research process itself. Every claim you encounter gets classified, scored, and triangulated as you go. The verification workflow integrates these steps into a repeatable pattern that becomes habitual with practice.

The workflow operates in three stages. Collect: gather the claim, note the source tier, tag the domain. Corroborate: seek at least one additional source from a different tier. Classify: assign a confidence level based on corroboration strength. Claims that pass all three stages go into your verified findings. Claims that fail at the corroboration stage go into your hypothesis file — they are not discarded, but they are clearly labeled as unconfirmed and earmarked for further collection.

1. Collect and Tag When you encounter a claim, immediately tag it with source tier, date, and domain. Do not evaluate it yet — just capture it with metadata. This takes five seconds and prevents the common failure of incorporating untagged information that later cannot be traced back to its origin.
2. Corroborate Within 24 Hours For any claim that might influence your conclusions, seek corroboration within one collection cycle. The faster you corroborate, the less likely you are to build analysis on unverified foundations. Even a quick search for a second source changes the reliability profile of a claim.
3. Classify and File Verified findings (triangulated, multi-tier) go into your working conclusions. Corroborated signals (two sources, same tier) go into your strong hypotheses. Uncorroborated claims go into your watch list. Everything is labeled. Nothing is lost. The classification determines how heavily you weight the information in your analysis.