Key Takeaways

  • Byron Ling has evaluated founders across roughly 30,000 meetings over the past decade.
  • Twelve Below deploys AI agents to scan digital writing breadcrumbs, surfacing founders before they publicly announce new companies.
  • Sourcing throughput has scaled, but competitive deal allocation still depends entirely on founder trust and qualitative judgment.
  • Early-stage firms are eliminating junior spreadsheet analysts in favor of mini partners who can build instant rapport.

Automated Breadcrumbs vs Human Deal Allocation

Venture capital has spent years treating sourcing as a manual grind. Associates clicked through LinkedIn updates, tracked alumni networks, and parsed corporate registry filings. That era is over. Byron Ling, partner at Twelve Below, relies on autonomous software to detect company formation signals months before founders launch publicly.

“The biggest thing so far has been increasing the throughput and quality of founders that we can meet,” Ling explains. “And a lot of this comes down to there's a lot of tools out there that can now identify individuals often before it's public that they're starting companies.”

Twelve Below runs AI agents across public digital breadcrumbs, such as technical blog posts, code commits, and subtle changes in online writing habits. This automated net widens the top of the funnel. Yet Ling draws a sharp boundary around where machine intelligence stops creating alpha.

Automating the discovery of an operator does not win the term sheet. “The parts that it's not going to replace is just the quality of your judgment,” Ling notes. “The quality of how you build trust in person. This is still a business where the founder picks the investor. It's not like buying stocks in the public markets.”

In private markets, access is an allocation battle, not a market order. If five funds identify the same repeat founder simultaneously, algorithmic detection gives zero edge. The win goes to the investor who forms an immediate human connection.

The Extinction of the Junior Modeler

Because software now handles outbound identification and initial data synthesis, the profile of the entry-level venture hire has changed. Early-stage seed investing rarely requires complex discounted cash flow spreadsheets. The traditional analyst role, centered on financial modeling and memo drafting, is dissolving.

Host David Weisburd pinpointed the structural shift inside fund rosters: “Respectfully, what you're really saying is that you've completely changed your hiring practice from these analysts that are basically doing this analysis and doing these models to what I would call mini partners, somebody that you want to see a partner five to 10 years out.”

Ling agrees that deal leads cannot outsource their core thinking to prompts or summaries. “The person who is the one building the relationship with the founder, who's the one who's going to eventually have to win, they have to be close to the work though,” Ling says. “I don't think they can entirely depend on AI producing the questions that need to be asked.”

Firms that rely on AI scripts to run diligence meetings end up looking generic. Founders detect when an investor is reading off an automated prompt. The edge at seed stage comes from rapid, intuitive pattern recognition built over thousands of hours in the room.

Why It Matters

AI commoditizes the discovery layer of private investing, driving the cost of finding deals toward zero. As deal discovery becomes table stakes, LP capital will flow to firms whose junior talent can win competitive founder bake-offs on relational credibility rather than administrative volume.