Executive Summary
Product development has fundamentally shifted from deterministic workflows to non-deterministic AI agent orchestration, creating massive disruption for legacy software companies. Gokul Rajaram, who built core ads systems at Google and Facebook and invested in 700+ companies, identifies a critical bifurcation in enterprise software survival. Companies pricing on seat-based utility face existential threats as AI agents can perform the same functions at fraction of the cost. Zendesk exemplifies this vulnerability - instead of 50 customer service seats at $20-30 each, companies can deploy 20 seats plus 30 AI agents. Meanwhile, system-of-record companies with timeless data remain protected. Salesforce and NetSuite benefit from career-limiting switching costs and deep data moats that require years of migration effort. The transformation accelerated dramatically in late 2025 when system-of-record companies began cutting API access to prevent AI companies from treating them as 'dumb databases.' Slack's blocking of Gleen represents the opening salvo. Legacy software companies must either go private for business model transformation or build outcome-based pricing. The winners will be AI-native companies that replace entire systems, not just workflows, and established platforms that successfully bundle AI agents with their data advantages.
Key Insights
what Gokul Rajaram said“The first thing we are seeing now happen is that PMs are starting to check in code with either the codex or the clod code into the actual production repository”
This is a preview. Log in to see the full analysis including investment opportunities, risks, catalysts, and detailed insights.