Your AI Pilots Won't Move the P&L. Here's What Will - Adam Nahari, Apollo
Summary
The speaker argues that while AI tools are widely adopted, the material impact on the Profit & Loss (P&L) is often negligible because organizations focus too heavily on technology adoption rather than fundamental process redesign and change management. The core thesis is that for every five units of change management, there is only one unit of engineering effort required for successful AI transformation. To achieve a 10x win, businesses must start by identifying a P&L-critical problem and redesigning the entire workflow around the new AI capability, rather than simply automating existing, inefficient processes.
Key takeaways
-
Focus on P&L, not technology, to drive AI value
9:05
Successful AI transformations must start by identifying a measurable problem that moves a line item on the P&L (e.g., EBITDA). The speaker notes that the biggest gains come from Core Operational AI (Layer 3), which is embedded in business processes and cannot be bought off the shelf. (2:42, 5:45)
-
The ratio of change management to engineering
4:02
The speaker posits that for every five units of change management, there is only one unit of engineering effort. The real effort lies in picking the right problem, creating buy-in, driving adoption, and measuring impact. (2:42)
-
The Six Moves for Successful AI Transformation
12:01
1. Pick one or two problems, starting with the P&L. 2. Reimagine the process end-to-end (the 10x play). 3. Start with the CEO (securing public commitment). 4. Build with the business (ensuring P&L owners and domain experts co-sponsor the change). 5. Put the tools in everyone's hands (requiring experimentation). 6. Set kill metrics first (using time, metric, and adoption gates). (7:21, 14:07)
Technical details
-
AI Value Layers
342s
AI value exists in three layers: 1) General Productivity (copilots/chatbots, gains show in price). 2) Domain-Specific AI (retrieval over enterprise documents). 3) Core Operational AI (vertical systems embedded in business processes, gains remain in margin). (3:42)
-
Optimization Engine Design
The speaker describes a 10x play in crew scheduling: building an optimization agent engine that consolidated company data (flights, crew, maintenance) and ranked possible solutions by cost, which was then integrated into a decision support app. (10:00)
-
Pilot Success Metrics
Before starting a pilot, three types of gates must be set: a time gate (e.g., 90 days), a metric gate (AB testing against a baseline), and an adoption gate (e.g., 50% usage rate). The CFO must also calculate the unit economic scalability. (12:40)
Mentioned resources
Channel & topics
Watch on YouTube · Back to latest
This independent, AI-assisted summary is provided for commentary and informational purposes. It may contain errors or omit important context. Please watch the original video for the creator's complete presentation. Video, thumbnail, and related copyrights belong to their respective owners.