How did the FDE model originate at Palantir?

The FDE model originated at Palantir out of necessity when building software for the intelligence community, where direct customer needs were hard to ascertain. The initial approach involved taking a demo to customers and iterating based on feedback, which evolved into the FDE strategy where engineers were embedded to discover and build solutions directly.

What's the difference between the FDE model and traditional consulting?

Unlike consulting, the FDE model aims to build repeatable, scalable value. While FDEs perform custom work, the goal is to generalize these solutions into core product features over time. It's characterized by growing contract sizes and increasing product leverage, rather than a focus on minimizing work per customer.

Why are so many AI startups adopting the FDE model?

AI agents represent a new market category with no incumbent products, necessitating significant product discovery. The FDE model allows startups to deeply understand customer needs from within the enterprise, build tailored solutions, and scale adoption in a rapidly evolving technological landscape.

How should startups price their solutions using the FDE model?

The FDE model shifts pricing from usage or subscription (SAS) to the value of the outcome delivered. Contracts tend to be larger and more flexible, focused on solving a specific problem and delivering quantifiable value to the customer, rather than a recurring fee based on software access.

What is the role of the product team in the FDE strategy?

The product team's role is to hold the overall product vision and identify generalizable solutions from the custom work done by FDEs. They must abstract the specific customer problem into a broader, scalable feature that can serve multiple customers, turning the 'gravel road' built by FDEs into a 'paved highway'.

How does the FDE model relate to YC's advice to 'do things that don't scale'?

The FDE model is effectively 'doing things don't scale at scale'. While early-stage advice suggests manual, high-touch efforts, the FDE approach applies this intensely customized, hands-on work to large contracts, enabling startups to go further by driving up contract value and leveraging product development.

What are the biggest opportunities for startup founders in AI right now?

Despite rapid AI capability improvements, adoption is lagging. The biggest opportunity lies in filling this gap: making AI capabilities genuinely useful to people by understanding and addressing customer needs from within the enterprise, which the FDE model is well-suited for.

The FDE Playbook for AI Startups with Bob McGrew

Y Combinator

Science & Technology5 min read51 min video

Sep 8, 2025|98,309 views|1,691|78

YC Y Combinator

Save to Pod

Key Moments

On this page

TL;DR

AI startups use the Forward Deployed Engineer model for product discovery and customer adoption.

Key Insights

The Forward Deployed Engineer (FDE) model involves technical staff embedded with customers to bridge product and user needs.

FDEs drive product discovery by solving unique customer problems, which informs future product development.

Unlike traditional sales, FDEs work from the 'inside' of the customer's organization, enabling deeper problem-solving.

The FDE model focuses on delivering specific outcomes and increasing contract value, rather than just scaling.

Successful FDE implementation requires careful team structuring (e.g., Echo and Delta teams) and specific hiring profiles.

AI agent startups are embracing the FDE model due to the lack of incumbent products and the need for extensive product discovery.

UNDERSTANDING THE FORWARD DEPLOYED ENGINEER (FDE) MODEL

The Forward Deployed Engineer (FDE) is a technical role, typically an engineer, embedded directly within a customer's. This individual acts as a crucial intermediary, identifying and filling the gap between the existing product's capabilities and the specific, often unarticulated, needs of the customer. The FDE's primary function is to solve unique problems for a given customer, thereby delivering significant value and enabling the product team to generalize these solutions for a broader market.

ORIGINS AND EVOLUTION AT PALANTIR

The FDE model originated at Palantir, born out of the necessity to build software for intelligence agencies where direct customer feedback was scarce and secretive. Initially, a demo-based approach was used, iterating based on feedback. However, it became clear that each customer's needs were subtly different, leading to the development of a platform adaptable to various sites. This shift necessitated embedding engineers who could customize the platform on-site, a role that was initially viewed as an unavoidable but undesirable service function.

THE FDE AS A PRODUCT DISCOVERY ENGINE

A key innovation of the FDE strategy, particularly championed by Palantir early employee Sham Sankar, was recognizing that these embedded engineers could be a powerful engine for product discovery. Instead of minimizing 'services' work, FDEs were tasked with actively identifying customer pain points and building custom solutions to meet those immediate needs. This 'gravel road' approach would then inform the product and engineering teams on how to build a generalized, scalable 'paved highway' for future customers.

STRUCTURAL AND TALENT CONSIDERATIONS FOR FDE TEAMS

At Palantir, the FDE model was supported by a structured team, often involving an 'Echo team' of embedded analysts who managed customer relationships and identified use cases, and a 'Delta team' of highly skilled software engineers who built and deployed solutions. Hiring for these roles required specific profiles: domain experts with a 'rebel' mindset for the Echo team to challenge existing norms, and prototyping-focused engineers for the Delta team, capable of rapid development rather than perfectionist coding. This blend proved highly effective in addressing intricate customer problems.

DIFFERENTIATING FDE FROM CONSULTING AND SaaS

The FDE model is distinct from traditional consulting, which often focuses on delivering services without a direct pathway to product generalization or scaling. While initially seeming like 'consulting dressed up,' the FDE model's value lies in its feedback loop to product development and its potential for profitable, repeatable solutions over time. Unlike SaaS, which aims for standardized, low-cost, scalable contracts, the FDE approach drives towards larger, more flexible contracts by delivering progressively greater outcomes and value.

THE AI AGENT BOOM AND THE FDE REVIVAL

The current AI agent boom has seen a resurgence of the FDE model, driven by the absence of established incumbent products. This market requires significant product discovery, which can best be achieved by embedding technical talent within customer organizations. The high speed of AI capability improvements outpaces adoption, creating a gap that FDEs are well-positioned to fill. Startups are leveraging this model to navigate new market categories and drive adoption, often finding that large contracts are feasible due to the inherent value and custom solutions provided.

PRICING BASED ON OUTCOMES, NOT USAGE

A critical aspect of the FDE model, especially for AI startups, is shifting from usage- or seat-based pricing (common in SaaS) to outcome-based pricing. The core offering is the successful resolution of a customer's problem, not just the installation of software. This means startups must be adept at defining, delivering, and monetizing the value of the outcomes they enable, often taking on more risk initially to demonstrate their execution capabilities and secure larger, evolving contracts as they expand within a client organization.

NAVIGATING EXECUTIVE BUY-IN AND ORGANIZATIONAL DISCIPLINE

Implementing the FDE model successfully requires significant organizational discipline to prevent it from devolving into pure, unscalable consulting. Securing executive buy-in is paramount, especially when dealing with large enterprises that may have entrenched IT processes or skepticism towards startups. The FDE must often navigate internal politics and convince stakeholders, particularly by aligning with the organization's top priorities, a task made easier when solving critical CEO-level problems, and demonstrating tangible, replicable value over time.

THE ROLE OF THE PRODUCT TEAM IN AN FDE ENVIRONMENT

The product team plays a vital role in generalizing solutions identified by FDEs. Instead of building highly verticalized products for single customers, product managers must possess the ability to abstract problems, identify commonalities, and design a core product or platform that is adaptable. This requires thinking at a higher level of generalization, ensuring that the product provides leverage to FDEs, enabling them to deliver increasing value to customers without proportional increases in customization effort. This creates a synergistic relationship where the product team supports and enables the FDEs.

DEMO-DRIVEN DEVELOPMENT AND LEARNING ORGANIZATIONS

The FDE model often leads to a form of 'demo-driven development.' By repeatedly showcasing solutions to new customers, teams are forced to refine their offerings and integrate new features seamlessly. This customer-centric approach drives product improvement by focusing on how features work together from the customer's perspective, creating desire and addressing real pain points. Ultimately, companies employing the FDE model must be learning organizations, continuously adapting and iterating based on field experience, much like a startup itself.

FUTURE OPPORTUNITIES IN AI ADOPTION

The rapid advancement of AI capabilities presents immense opportunities, but adoption often lags. There's a significant gap between what AI can theoretically do and what is practically implemented and useful for customers. Startup founders can capitalize on this by focusing on filling this adoption gap, much like FDEs fill the gap between products and customer needs. This involves human ingenuity, exploration, and a willingness to tackle the 'pain' of integration to make powerful AI capabilities genuinely accessible and beneficial to users.

Mentioned in This Episode

●Software & Apps

●Companies

●Organizations

●Books

●Concepts

●People Referenced

Forward Deployed Engineer (FDE) Playbook for AI Startups

Practical takeaways from this episode

Do This

Act as product discovery agents, filling the gap between product and customer needs.

Focus on solving impactful problems that are top priorities for leadership.

Embrace the 'land and expand' model, identifying and solving progressively more valuable problems.

Structure teams with 'echo' (embedded analysts) and 'delta' (software engineers) roles.

Hire 'rebels' or 'heretics' for echo roles who recognize current insufficiencies.

Seek prototyping experts for delta roles who can deliver rough-and-ready code quickly.

Price based on the outcome and value delivered, not usage or subscription.

Drive contract size up by delivering increasingly valuable work.

Measure success by the value of the outcome delivered and increasing product leverage.

Build a learning company, fostering continuous adaptation and improvement.

Focus on the customer's perspective and desires, using demos to create desire.

Understand that the FDE model requires discipline and is a skill that can be learned.

Avoid This

Do not treat FDE as simply consulting; focus on building repeatable value.

Avoid building products that are over-specialized for a single customer.

Don't solely rely on sales-led product discovery; engineers should be involved.

Do not aim to minimize customer engagement; deep customer understanding is key.

Avoid hiring traditional salespeople for government/defense contracts; culture fit and technical understanding are crucial.

Do not build software that is difficult for FTEs to use, hindering their leverage.

Refrain from blindly implementing what customers ask for if it's not impactful.

Do not treat FDE as a 'do this at home' strategy unless it's the only way to succeed in your market.

Avoid comparing FDE too closely to standard SaaS models; pricing and contract structures differ.

Don't neglect the IT department or internal gatekeepers when deploying on-premise solutions.

Avoid becoming complacent; continue learning and adapting, even as a large company.

Common Questions

A Forward Deployed Engineer (FDE) is a technical professional, typically an engineer, who works directly at a customer's site. Their primary role is to bridge the gap between the existing product capabilities and the specific needs of the customer, often discovering and implementing solutions to unique problems.

Topics

FDE Model Product Discovery Customer Engagement US Army Reserve

Mentioned in this video

conceptPalantir ontology

The foundational data structure that allows for generalization and customization across different customers and verticals within Palantir's government product.

personGeneral Randy George

Chief of Staff of the Army who has articulated a plan for the army's transformation.

organizationUS Army

Bob McGrew has an new role exploring the future of AI with the US Army.

personSham Sankar

An early employee at Palantir, credited with inventing the FDE strategy by realizing FTEs could act as product discovery agents.

organizationUS Army Reserve

Bob McGrew's new role within this organization involves advising on technology.

conceptForward Deployed Engineer (FDE)

A technical role that sits at the customer site to bridge the gap between product capabilities and customer needs, often driving product discovery and customization.

softwareGBD40

Mentioned as a recent AI model release (April 2024) indicating rapid capability improvements.

toolCastle