# Why Buying Companies Beats Selling Them AI — Prathik Naidu, Long Lake

## Executive summary

Prathik Naidu argues that deploying AI into enterprises is significantly more challenging when approached as an external vendor ('outside-in'). He proposes that the superior strategy is acquiring established service businesses and transforming them internally ('inside-out'). This model leverages existing trust, deep operational knowledge, and proprietary data to achieve true P&L impact, especially in the services economy, where AI can create operating leverage by reducing labor costs.

## Key takeaways

- Challenges of Outside-In AI Deployment: Deploying AI as a third-party vendor faces hurdles including navigating complex organizational structures, understanding 'tacit knowledge' (human judgment and edge cases), and difficulty capturing enduring business value (e.g., distinguishing a $1M contract from a potential $5M EBITDA impact).
- The Value of Ownership and Integration: Ownership provides critical advantages: deep access to operational teams and veterans (20 years of experience), access to a live data environment of real tasks and edge cases, and superior incentive alignment, ensuring the team's success is tied directly to the business's P&L.
- AI's Impact on Service Economics: Services businesses traditionally face high marginal costs (e.g., 80% overhead for every $100 of revenue). AI deployment, however, can introduce operating leverage, allowing service economics to resemble software models by improving quality while reducing direct labor cost scaling.

## Technical details

- Operational Complexity: AI agents must integrate with systems of record (CRM, ticketing systems, document repositories) but must ultimately solve for 'human judgment' and 'tacit knowledge'—the exceptions and edge cases that govern business operations.
- Value Capture and Economics: The true value of an AI deployment is measured by its impact on the P&L of the business, not just the contract revenue. The speaker notes that a conservative 10x multiple applied to generated EBITDA impact can significantly increase enterprise value.
- Transformation Curve (J-Curve): The transformation process involves an initial investment phase (building the platform, capturing workflows, training models) followed by a steep upward curve where capacity and time savings are unlocked, culminating in measurable P&L impact and improved service quality.

## Practical implications

- For build engineers, the focus must shift from merely deploying a tool (AI agent) to engineering for deep operational integration and measurable P&L outcomes.
- The most valuable data sources are not structured records, but the 'edge cases' and workflows derived from expert human judgment.
- AI solutions must be designed to solve for 'operating leverage' in service businesses, ensuring that growth in revenue does not require a proportional, linear increase in labor costs.

## Topics

AI Deployment, Service Economy, M&A Strategy, Operational Excellence, P&L Transformation, AIE CODE summit, Long Lake, Amex GBT

Source: https://www.youtube.com/watch?v=J7n3WJnFTto