The FDE Playbook for AI Startups with Bob McGrew

Full Title

Summary

This episode explores the Forward-Deployed Engineer (FDE) model, a strategy pioneered by Palantir for customer-centric product development and adoption, and its relevance to the current AI startup landscape.

The discussion highlights how the FDE model addresses the unique challenges of deploying novel AI technologies by embedding engineers with customers to drive product discovery and build tailored solutions, contrasting it with traditional SaaS scaling.

Key Points

The FDE model involves technical engineers embedding with customers to bridge the gap between product capabilities and specific customer needs, often solving novel problems.
Palantir's FDE strategy originated from the challenge of building software for the intelligence community, where direct user feedback was scarce and conventional sales approaches were ineffective.
The FDE model prioritizes deep customer engagement and problem-solving over generalized scaling, allowing for the creation of platform-like solutions adaptable to diverse customer environments.
The distinction between sales-led and FDE-led product discovery is crucial, with FDEs gaining insights from within the customer's operations, leading to more impactful solutions.
Palantir's FDE team was structured with "Echo" (analyst/account management) and "Delta" (software engineering) roles, requiring unique profiles focused on domain expertise, critical thinking, and rapid prototyping.
The FDE model can initially result in losses but becomes profitable as the product becomes better suited to customer needs and expands to address more significant problems, demonstrating a scalable approach to complex enterprise solutions.
The inherent heterogeneity of markets, like government and defense, necessitates the FDE approach, where solutions are tailored to specific segments rather than a single, standardized product.
The rise of AI agents is driving the adoption of the FDE model because there is no incumbent product, requiring extensive product discovery from within the customer's context.
The FDE approach is effectively "doing things that don't scale at scale," leveraging deep customer engagement to drive adoption of novel technologies.
Misconceptions about the FDE model often equate it to consulting; however, successful FDE implementation focuses on building repeatable value and driving product generalization.
Pricing for FDEs shifts from usage-based SaaS models to outcome-based contracts, reflecting the value delivered to the customer and encouraging larger, more flexible agreements.
Building a successful FDE strategy requires significant organizational discipline, a focus on delivering valuable outcomes, and a commitment to continuous learning and adaptation.
The FDE model is well-suited for AI startups because the AI landscape is rapidly evolving with no established products, demanding intensive, customer-specific product discovery.
Demo-driven development is critical for the FDE model, forcing a customer-centric perspective and driving the creation of integrated, valuable solutions that solve real pain points.
The FDE model fosters a "learning company" culture, essential for adapting to rapidly evolving AI capabilities and ensuring product adoption.

Conclusion

The Forward-Deployed Engineer (FDE) model is crucial for the current AI landscape due to the lack of incumbent products and the need for deep customer engagement to drive adoption.

Building a successful FDE strategy requires significant organizational discipline, a focus on outcome-based pricing, and continuous learning, akin to operating as a startup.

The FDE approach embodies Y Combinator's ethos of "doing things that don't scale" but at a scale that is now feasible with large enterprise contracts, filling the gap between AI capabilities and their practical application.