The FDE Playbook for AI Startups with Bob McGrew

McGrew describes an FDE as a technical person who works at a customer site to bridge the gap between a product and what the customer needs. At Palantir, teams brought an early demo to intelligence customers, gathered feedback, and built solutions for problems they had not encountered before.

Palantir found that customers needed slightly different products, so it built a customizable platform rather than separate products for each site. FDEs created a working solution, or “gravel road,” while the product team generalized it into reusable software for other customers. McGrew says the model must produce repeatable value and lower the cost of delivering outcomes over time to avoid becoming consulting.

Palantir’s Echo team worked with users and managed customer relationships, while its Delta team built and deployed prototypes quickly. McGrew says Echo members needed domain knowledge and a willingness to challenge existing practices; Deltas needed to be strong at rapid prototyping. Product teams then worked with FDEs across several customers to find abstractions that could serve more than one workflow.

McGrew links FDE hiring to AI agents being a new market category with no incumbent product and substantial product discovery still to do. He says this discovery often requires working inside enterprises, where needs vary across market segments. For AI startups, he recommends selling outcomes, securing executive backing, and tracking both the value delivered and the product leverage gained.

McGrew describes joining the U.S. Army Reserve as an officer advising the Army on technology, with authority to work on problems from the ground up and escalate them to leadership. He compares parts of that work to FDE practice, including connecting leadership priorities with on-the-ground needs. He sees a startup opportunity in closing the gap between rapidly improving AI capabilities and what customers can actually adopt.