The FDE Playbook for AI Startups with Bob McGrew

Key Concepts

• Forward Deployed Engineer (FDE): A technical engineer who works on-site with customers to bridge the gap between a product's capabilities and the customer's specific needs.
• Echo Team: Embedded analysts at the customer site who understand user needs and manage relationships.
• Delta Team: Deployed engineers who rapidly prototype and build solutions based on the echo team's insights.
• Product Discovery: The process of identifying and understanding customer needs to inform product development.
• Land and Expand: A sales strategy focused on initially solving a key problem for a customer and then expanding to address other valuable problems within the enterprise.
• Ontology: A structured representation of knowledge as a set of concepts within a domain and the relationships between those concepts.
• Product Leverage: The ability of a product to deliver increasing value to customers with less customization or effort.
• Outcome-Based Pricing: Pricing a solution based on the value it delivers to the customer, rather than usage or subscription.
• Demo-Driven Development: A product development approach focused on creating compelling demos that showcase the product's capabilities and create customer desire.

Palantir's Forward Deployed Engineer (FDE) Model

Origins and Motivation

• Palantir developed the FDE model because its initial focus was building software for the intelligence community, where understanding customer needs was challenging due to secrecy and the unique requirements of each agency.
• The traditional approach of achieving product-market fit and then scaling by distancing from customers didn't work for Palantir because each customer's needs were slightly different.
• Sham Sankar, a key figure at Palantir, realized that FDEs could act as product discovery agents, bridging the gap between the product and user needs.

How the FDE Model Works

• Team Structure:
  • Echo Team: Embedded analysts who understand user needs, manage relationships, and identify key problems to solve. They need deep domain knowledge and a "rebel" or "heretic" mindset to recognize the limitations of existing solutions.
  • Delta Team: Deployed engineers who rapidly prototype and build solutions based on the echo team's insights. They need to be skilled at writing code quickly and delivering outcomes on a timeline, even if the initial code needs to be rewritten.
• Process:
  1. The echo team identifies a key problem for the customer.
  2. The delta team builds a prototype solution.
  3. The solution is presented to leadership for feedback.
  4. If successful, the solution is deployed organization-wide.
• Key Principle: The FDEs "build a gravel road" to where the product needs to go, and the product team then turns that into a "paved superhighway" for broader applicability.

Distinguishing FDE from Traditional Consulting

• The FDE model is not just consulting dressed up with fancy marketing.
• In a successful FDE deployment, the product improves over time, reducing the need for a large on-site team.
• The FDE team earns the right to address more important problems at the customer site, increasing the value delivered.
• Profit margins start negative but become positive as the product becomes better suited to the customer's needs.

The Role of the Product Team

• The product team's role is to hold the product vision and generalize solutions for broader applicability.
• They need to identify the underlying problem being solved and create a generalizable version that applies to multiple customers.
• The product team should focus on building a platform that can be customized by FDEs, rather than building highly verticalized solutions.
• The Palantir ontology, which allows for defining objects, properties, and relationships, is a prime example of a generalized solution.

Managing Tension and Maintaining Discipline

• There is inherent tension between the FDEs, who want to build specialized solutions for individual customers, and the product team, who want to build generalizable solutions.
• Organizational discipline is crucial to prevent the FDE model from devolving into pure consulting.
• It's important to avoid building products in the field that customers are merely asking for, rather than those that are truly valuable to them.

The Rise of FDEs in AI Agent Companies

Why FDEs are Relevant in the AI Agent Space

• The AI agent market is characterized by significant heterogeneity and a lack of established solutions.
• AI agents represent a completely new market category, unlike incremental improvements to existing SAS products.
• There is a significant amount of product discovery to be done, which can only be accomplished from within the enterprise.
• "With AI agents, there is no incumbent product."

Pricing and Contract Considerations

• FDE-driven companies should focus on outcome-based pricing, rather than usage-based or subscription-based pricing.
• Contracts should be larger and more flexible, reflecting the complexity and value of the solutions being delivered.
• Startups should be willing to take on risk early on, believing in their ability to execute and deliver results.

Overcoming Enterprise Challenges

• Startups need to navigate internal politics and gain executive buy-in to succeed in large enterprises.
• It's crucial to work on one of the CEO's top five priorities to ensure the project receives the necessary attention and resources.
• Startups may need to challenge established IT practices and advocate for the flexibility needed to deploy their solutions.

Scaling and Abstraction

• While it's okay to do custom work per customer initially, the goal should be to get less custom per every customer over time.
• The key metric is the value of the outcome being delivered, not necessarily the amount of custom work being done.
• The product should provide increasing leverage to the FDEs, enabling them to deliver more value with less effort.

Demo-Driven Development in the FDE Context

• Demo-driven development works well in the FDE context because it forces companies to repeatedly showcase their solutions to new customers.
• A good demo creates desire in the customer for what the company is doing, indicating that a real pain point has been identified.
• Demos also force companies to think about how different features work together and create a seamless user experience.

Building a Learning Company

• The FDE model requires building a learning company that is constantly adapting and improving.
• Companies should focus on filling the gap between the capabilities of AI and the ability of customers to adopt them.
• "The FD model effectively is doing things that don't scale at scale."

Bob McGrew's New Role with the US Army Reserve

Joining Detachment 2011

• Bob McGrew has joined the US Army Reserve as a lieutenant colonel in Detachment 2011.
• The unit advises the army on technology and is composed of officers who have taken the oath and undergone military training.
• The goal is to help the army transform from fighting the wars of the past to preparing for future conflicts.

Applying Palantir Lessons to the Army

• McGrew plans to apply the FDE strategy to the army, identifying the leadership's top priorities and working directly with officers on the ground to solve problems.
• He aims to help bridge the gap between the army's goals and the existing implementation practices.

Opportunities for Startups

• McGrew believes that the best opportunities for startups lie in filling the gap between the capabilities of AI and the ability of customers to adopt them.
• He sees a need for human ingenuity and exploration to make AI genuinely useful to people.

Conclusion

The FDE model, pioneered by Palantir, is a powerful approach for companies operating in complex and heterogeneous markets. It emphasizes close collaboration with customers, rapid prototyping, and a focus on delivering valuable outcomes. While it requires organizational discipline and a willingness to challenge conventional thinking, the FDE model can be a significant competitive advantage, particularly in emerging fields like AI agents. The key is to build a learning company that is constantly adapting and improving its solutions based on real-world customer needs.