From Sales Engineer to Forward Deployed Engineer

This series has argued that selling AI is different (post 1): the demo is a trap, the POC is the crux, and the technical win is earned on real data. All of that points at one conclusion about who wins these deals.

The spectrum

Technical go-to-market has always run along a spectrum:

• Sales Engineer (pre-sales) — sells the vision: discovery, demo, proposal, technical evaluation. Hands off to implementation at the close.
• Forward Deployed Engineer (deploy-by-building) — embeds with the customer and proves value by building it: bespoke integrations, prompts, agents, pipelines, on their real data.
• Implementation / Customer Success Engineer (post-sales) — makes it stick and expands.

The classic motion put the weight on the left: win the deal in pre-sales, then throw it over the wall. AI shifts the weight toward the middle — because, as this series has shown, value in AI isn’t credible until it’s built and proven in the customer’s environment. The hard, deal-deciding work moves from presenting to building, and the forward deployed engineer — pioneered at Palantir, now the signature GTM role at AI companies — becomes the center of the sale.

Why AI made the FDE mandatory

The forward deployed engineer isn’t new; the AI inflection made it necessary. Three forces:

• Demo ≠ value. When the demo can’t prove it works for them, someone has to actually make it work for them to close the deal. That someone is an engineer, embedded.
• The gap is too wide to self-serve. The distance from a capable model to ROI in production — real data, cost, security, edge cases, trust — is more than a customer can reliably cross alone. Put an engineer beside them and the gap closes.
• Trust is built by building. An anxious AI buyer trusts what they’ve watched work on their own data, run partly by their own people. That trust is a byproduct of embedded building, not of a great pitch.

For an engineer, this is the good news: the scarcest, least-fakeable asset in AI sales — the ability to actually build the thing — is now the front of the sale, not a backstage skill. “The engineer who closes by building” is a real and defensible identity.

The craft tension — and the discipline that resolves it

There’s a danger in all this. A forward-deployment motion can rot into an unscalable custom-build sweatshop: every deal a bespoke project, nothing reusable, margins and sanity eroding. The craft that prevents it is scoping and productizing the learnings — turning each bespoke win into repeatable pattern and platform. That’s the fifth stage of the Deploy-to-Win Loop (productize), and it’s what separates a forward-deployment practice that scales from one that burns out.

This is also where the technical-GTM spectrum becomes a career and a business, not just a role — the build → scale → run arc of the Forward Deployment Stack: build the first win as an FDE, make it repeatable as an architect, run it as a function as a leader.

Where this series leads

If you’re an engineer, this is an invitation: the fundamentals of technical sales still matter (start with the Sales for Engineers series), and the AI-era patch is what this series added — but the role itself is tilting toward builders. The deeper technical craft of doing the work lives in The AI Forward Deployed Engineer; the craft and business of doing it as a practice lives in The Forward Deployed Consultant.

Key takeaways

• Technical GTM runs a spectrum — Sales Engineer (pre-sales, sells the vision) → Forward Deployed Engineer (deploy-by-building, proves value on real data) → Implementation/CS (post-sales) — and AI shifts the weight toward the middle because value isn’t credible until it’s built and proven in the customer’s environment.
• AI made the FDE mandatory: demo ≠ value (someone must make it work for them to close), the model-to-production gap is too wide to self-serve, and trust is built by building (anxious buyers trust what they’ve watched work on their own data).
• For engineers this is good news: the ability to build — the least-fakeable asset in AI sales — is now the front of the sale, making “the engineer who closes by building” a real, defensible identity.
• The craft tension: a forward-deployment motion can rot into an unscalable custom-build shop; the discipline that resolves it is scoping + productizing the learnings (the productize stage of the Deploy-to-Win Loop) — turning bespoke wins into reusable platform.
• The spectrum becomes a career/business via the build → scale → run arc (Forward Deployment Stack); the whole series’ point is that AI changed where deals are won (demo → production) and therefore who wins them — the forward deployed engineer is the apex of AI-era technical GTM.

Further reading

• Sales engineering — the pre-sales role at one end of the spectrum
• Solution selling — the outcome-first mindset the forward-deployed role takes to its conclusion