Gabe Stengel - Building Investing Superintelligence - [Invest Like the Best, EP.492]

Early applied AI companies struggled because underlying models weren't powerful enough, leading to poor user experiences. Winners built for the future capabilities of models, and their products became "magical" only when the technology caught up, validating their long-term vision.

Rogo found its best initial market in private market "dealmakers," not public equities. Private markets are full of manual, human-driven workflows ("plumbing") that AI can automate, representing a larger immediate business opportunity than simply analyzing already-available public data.

In finance, a user cannot trust an AI's output without understanding its origins. An answer that is auditable—showing data sources and assumptions—is more valuable and actionable than a "black box" answer because it allows users to verify and debug the process.

Companies like Rogo compete with OpenAI not by building better models, but by building perpendicular to them. They focus on complex, industry-specific "plumbing" like compliance systems, audit trails, and data rooms—critical infrastructure that is too niche for large horizontal players to prioritize.

Seat-based and token-based pricing for AI are intermediate steps. The ultimate model is outcome-based, where customers pay for specific value delivered—like a successful investment idea or a completed report. This perfectly aligns vendor cost with customer value, bypassing debates over token consumption ROI.

For decades, the value of investment firms was concentrated in their human talent. AI fundamentally shifts this, moving enterprise value towards proprietary software, data, and systems. This creates an existential threat for incumbents who must now compete with new, asset-light, AI-native firms.

To solve the challenge of rapid scaling, Rogo created an internal AI system that records all company conversations and knowledge. This "company brain" acts as an enablement tool, allowing new hires to quickly access historical context and expertise, dramatically reducing ramp-up time for roles like enterprise sales.

The current AI market is a land grab. The optimal strategy, advised by VCs like Sequoia's Pat Grady, is to be hyper-aggressive. Founders should accept a higher chance of total failure if it also increases the probability of achieving a massive, category-defining outcome, as the market will consolidate quickly.

As AI handles the analytical heavy lifting, the most valuable skill for investment professionals becomes gathering proprietary data. This means spending time in the field, speaking to experts, and building unique relationship graphs to feed their models with exclusive inputs not available to others.