Claude deep-research replaces $20 Perplexity (founder-level TAM research)

A 4-hour founder-level research workflow that replaces the $20 / month Perplexity habit. Two halves: (1) 2 hours of Claude deep-research (a stack like Claude's research mode + Parallel Search MCP + curated web), gathering citations and quotes; (2) 2 hours of deliberately loading the citations into the founder's head — reading the sources, building intuition, marking where the research is thin. Output: one slide of TAM at the intersection of 3 niches, with citations + quotes from named people + references to public initiatives, that holds up to VC scrutiny.

A small franchise founder built the slide that anchored his VC pitch in 4 hours. Same founder previously spent multiple weeks + a paid consultant on the same artifact. The discipline that matters most is the 2-hour "absorption" pass — without it the citations are a fragile mosaic that breaks on the first sharp VC question.

1. Define the question: usually "TAM × intersection of 3 niches" with a target customer profile.
2. Frame the research brief: what counts as evidence (named individuals, public initiatives, government data, segment reports).
3. Run Claude deep-research with the brief. Save every cited URL.
4. Run Parallel Search MCP / web search to verify 5–10 surprising claims with a second source.
5. Absorption pass (2 hours): open every primary citation, read for 3–10 minutes each, take notes in your own words. Flag where the original source is weaker than the synthesis suggests.
6. Compose one slide: number + 3 supporting bullets + 2 named quotes + 1 surprising fact. Cite inline.
7. VC-test: show the slide to one trusted operator before the actual VC meeting; revise.

• Skip the 2-hour absorption pass and the slide collapses on the first VC question. The research surface is brittle without the founder's brain behind it; that's the unique value.