Step 9: Validate and verify your research
Phase 1: Preliminary Research · Step 9 of 13
AI produces plausible content regardless of accuracy. Before you build anything on top of this research you need to verify the claims that matter most.
Article
This step exists because AI is confidently wrong often enough to be dangerous. Large language models generate plausible text, not verified truth. Every output from the previous steps contains a mix of accurate information, reasonable estimates, and potential fabrications. Your job is to figure out which is which before you build strategy on top of it.
The good news is that you do not need to verify everything. You need to verify the claims that your strategy depends on. If your entire business case rests on the market growing at twenty percent annually, that growth rate needs to be verified against a primary source. If a supporting detail about a competitor's founding date is off by a year, it does not matter.
Verification has a hierarchy. Primary sources such as government data, SEC filings, and company financial reports are the gold standard. Secondary sources such as analyst reports and quality journalism are reliable but may contain interpretation. Tertiary sources such as blog posts, general overviews, and unsourced claims should be treated as hypotheses until confirmed.
The most dangerous claims are the ones that feel obviously true. These are the claims you are least likely to check and most likely to build on. Market size figures, growth rates, and competitive positioning claims all fall into this category. They sound authoritative when generated by AI but may be assembled from outdated, misinterpreted, or entirely fabricated sources.
Your output is an annotated version of your research that distinguishes between what you know, what you estimate, and what you assume. This distinction is not academic. It determines how much confidence you should place in every subsequent decision.
Key Takeaways
- Why: A flawed assumption validated by a survey is still flawed. It just feels more certain.
- What: Review your research outputs and verify the three to five claims most critical to your business hypothesis.
- Output: An annotated research summary distinguishing verified facts from estimates from hypotheses.
AI Prompt
Act as a fact-checker and research validator. Here is my research summary from the previous steps: [paste your outputs from steps 2 through 8 here].
Your job is to help me identify what needs verification before I build strategy on top of this research. For each major claim in my research:
- Categorize it as verified fact, reasonable estimate, or unverified hypothesis
- Identify the original source if one was cited and note whether it appears credible and current
- Flag any claim that would significantly affect my business strategy if it turned out to be wrong
- Suggest the fastest way to verify the three most critical unverified claims
Do not verify claims yourself. Help me understand what I need to go check manually and why it matters.