Key Takeaways

  • Ninety-five percent of enterprise AI pilots fail to reach production because pure technical tools fail to address legacy workflows and regulatory constraints.
  • Silicon Valley has tried to engineer salespeople out of enterprise software since 2008, yet complex enterprise deals still require human trust.
  • Joseph Coyne targets what he terms an unfair fight: pairing fast AI development with operators who lived inside target customer organizations for 20 years.
  • Horizontal AI companies function primarily as utility layers, while defensible software margins belong to vertical products in esoteric sectors like insurance underwriting, lending, and airlines.

The Fallacy of Pure Technical Superiority

David Weisburd traces a familiar cycle back to 2008, when he entered venture capital. Software engineers have spent almost two decades attempting to code away the sales function. In enterprise software, technical founders routinely assume superior model performance or clean architecture will sell itself.

Enterprise reality rejects that assumption. Ninety-five percent of enterprise AI pilots fail. The bottleneck is rarely compute speed, parameter counts, or algorithm design. The breakdown happens at the interface between the software and legacy workflows in regulated verticals.

Joseph Coyne argues that horizontal software operates under a different set of economics: “When we think about horizontal, we think about more utilities that are foundational to AI.” (play from 39:24) Foundation models and horizontal developer tools build generic capabilities. Yet when software touches esoteric operational environments, such as commercial lending, insurance underwriting, or airline operations, generic technical excellence loses.

In these environments, buyers do not purchase models. They purchase risk mitigation, compliance coverage, and continuity with decades of existing enterprise procedures. A technical founder pitching a 5% accuracy gain will lose to a solution that integrates with legacy systems and meets internal risk committee demands.

Engineering an Unfair Fight in Regulated Sectors

Enterprise buyers run on trust. A chief risk officer at a Tier 1 bank or an operations executive at a major airline does not buy software from someone who treats their sector as an implementation detail.

“The best engineers, unless they're the founder of a company, are not going out and talking to the customer or have lived inside the customer for 20 years,” (play from 42:42) Coyne notes. Software engineers excel at building systems quickly, but they rarely understand the unwritten rules, internal politics, or regulatory traps inside legacy buyers.

Studio VC seeks investments structured around structural distribution advantages. As Coyne puts it: “And so when we're looking at a company, we're looking for an unfair fight. bringing a gun to a knife fight, right? It's who knows this industry better, who's better connected to this industry, who has the license to win in this industry.” (play from 41:07)

Selling into an esoteric vertical requires insider capital. That capital takes the form of personal relationships, pre-existing compliance credibility, and an intimate grasp of how budget allocations clear audit committees. An engineer paired with an industry veteran who spent two decades inside the customer's executive suite creates a distribution wedge that pure engineering startups cannot match.

Why It Matters

Technical moats in software are compressing as foundation model capabilities turn into utility infrastructure. Value accrues to vertical application layers that control proprietary customer workflows, data behind corporate firewalls, and direct executive distribution. Underwriting enterprise software investments requires prioritizing domain ownership and commercial distribution over raw model benchmarks.