Tanya M. Lawson Work With Me

HOW I RESEARCH

Not the loudest answer — the checked one.

When the stakes are high, the first answer is rarely the full answer. I research by looking for the missing pieces: what a claim assumes, who benefits if I believe it, and what evidence actually backs it up. That applies to expert opinions, official-sounding documents, AI outputs — and my own memory.

What this page is

How I separate signal from noise

Research isn't collecting information. It's deciding what deserves trust. Here's the discipline I use to make that call.

You’ll see
  • separating fact from opinion — even when opinion sounds authoritative
  • spotting conflicts of interest behind “expert” advice
  • cross-checking sources instead of trusting one voice
  • using AI as support — not as a truth machine
Bottom line
  • Proof before trust. A claim earns belief by surviving a check — not by sounding confident.

The Claim Check Method

A simple discipline for deciding what deserves trust.
The method

Five questions before I trust a claim

  1. Where did this come from? A named source I can check, or “somebody said”?
  2. What is it assuming? Most bad decisions hide inside unexamined assumptions — I list them out. I call this part the Assumption Audit.
  3. Who benefits if I believe it? Incentives shape advice more than people admit.
  4. What would prove it wrong? If nothing could, that's a belief — not a fact.
  5. Does a second, independent source agree? One source is a lead. Two independent sources are evidence.

I apply the same check to narratives, AI outputs, my own assumptions, and even my own memory of events. Confidence is not evidence — records are.

Where this matters most
High-stakes questions.

• legal and compliance research
• financial and licensing decisions
• vendor and tool selection
• anything an AI tells you with a straight face

AI, specifically
Support, not authority.

AI is excellent at gathering, summarizing, and drafting — and confidently wrong often enough that I never treat its output as a finished fact. AI answers go through the same claim check as everything else before they touch a real decision.

Why it works
It catches the errors that cost the most.

Most expensive mistakes don't come from missing information — they come from trusting the wrong information early and building on top of it. Checking claims up front is slower for an hour and faster for a year.

Facing a decision built on unchecked claims?

I’ll help you find out what's actually true first.

Tell me the decision and what it currently rests on. I'll help you check the claims, surface the assumptions, and document what the evidence really supports.

Work With Me Back to How I Work