How to Find Family Office… | The VC Corner Summary

1. Key Themes

Family offices are undervalued because they're hard to find, not because they're unwilling to invest

The article frames access difficulty as a moat rather than a deterrent.

"That fourth row is the whole opportunity. Difficulty of access is what keeps the competition out, and every founder who gives up at that row leaves the office to whoever kept going."

Most founders fail before they even start sourcing — by treating "family office" as a monolith

The critical error happens in the framing/definition stage, not the outreach stage.

"The biggest mistake happens before any tool gets opened: people look for 'family offices' as if that were one type of investor. It covers single family offices, multi family offices, family holding companies investing directly, offices that only go through funds, offices that anchor emerging managers, and offices doing direct deals in one sector because the family made its money there."

AI tools (specifically Claude) can compress days of manual research into an afternoon

The piece positions LLM-assisted sourcing as a structural time/cost advantage for fundraising founders.

"The offices are findable. The work is different from VC sourcing, and once the loop is built it runs in an afternoon." "Twenty minutes here beats a week of sending."

Personalization at scale is the new outreach standard

The playbook's later (paywalled) sections emphasize individualized messaging and enrichment rather than generic blasts.

"outreach where every message is written for one person" "The enrichment pass, so every contact arrives with two or three facts attached before you write a word"

2. Contrarian Perspectives

Going back to VCs who already passed is worse than the "harder" path of chasing invisible family offices

Conventional fundraising wisdom might favor maximizing volume with known, listed investors (VCs), but the author argues this is the inferior strategy compared to pursuing harder-to-reach family offices.

"Which is why most founders give up on them and go back to spraying VCs who have already passed."

The first list of investor categories you generate is worthless — real signal starts on the second and third pass

Rather than treating an initial AI-generated list as sufficient, the article insists the obvious answers are noise, and the value is in iterative refinement.

"The list of investor types roughly doubles past the first answer. The first answer is the same ten categories everybody gets, and the second and third pushes are where the offices nobody else is writing to show up."

3. Companies Identified

Claude (Anthropic) — AI assistant/LLM
- Why mentioned: Used as the core tool for generating and expanding lists of plausible family office investor types from a founder's deck
- Quote: "Give Claude your deck, your thesis, as much context as you can, and ask who could plausibly write this cheque."
HeyReach — Outreach/automation tool
- Why mentioned: Referenced as the platform for sending personalized messages at scale as part of the sourcing pipeline
- Quote: "The sheet-to-HeyReach pipeline that sends a unique researched message to every person on the list, at volume, with zero copies"

4. People Identified

Ruben Dominguez — Author/writer at The VC Corner
- Why mentioned: Byline author of the newsletter article and playbook
- Quote: (listed as author) "Ruben Dominguez... Sep 7"

5. Operating Insights

Reframe your target list before searching, not after. Use AI to exhaustively map investor sub-types (direct-investing families, fund-of-funds-style offices, sector-specific family money, emerging-manager anchors) before writing a single outreach message — this determines message strategy downstream.

"Each is a different search with a different message."
Push AI tools past their first output. Treat the initial list as a floor, not a ceiling — explicitly prompt for broader categories, narrower categories, missed regions, and alternative investment structures (direct vs. fund-based).

"Then push it. It gives you ten suggestions, and you say okay, what else? What is a broader way to cut this? What is a narrower one?"
Rank targets by reply-likelihood, not just relevance. The suggested Claude prompt explicitly asks for a ranked list by response probability, treating cold-outreach yield as a quantifiable, optimizable variable.

"Then rank the full list by how likely they are to answer a cold message."

6. Overlooked Insights

Your own network (LinkedIn) may already contain warm paths to these "invisible" investors — a potentially higher-signal shortcut than cold outreach entirely, mentioned only in passing as a forthcoming section.

"The warm-intro map. Turning your own LinkedIn export into a ranked list of who already knows your targets, with the exact prompt and the one-line ask"
There are non-outreach "back doors" into the most unresponsive family offices, implying that cold messaging isn't always the final answer even with perfect targeting and personalization.

"Plus the two back doors into offices that answer nothing cold"