In 2026, the Forward Deployed Engineer (FDE) has become what the ICO was in 2017 or the SPAC in 2021: a term few can define, yet everyone finds provocative. They seem to be the driver of the new AI economy. And Palantir is at the forefront of the revolution. But not everyone is falling for the magic.

A few weeks back I wrote about the Faculty's acquisition by Accenture, and received a very interesting comment on LinkedIn with regards to the talent angle of the deal: 

And if that’s not direct enough, you can always rely on Reddit to cut right to the chase: FDEs are just a new reincarnation of the good old deployment consultant. 

It is, I'll admit, a grounded take. Would adopting Palantir's org structure and titles guarantee a $350bn market cap and a ~230 P/E ratio? Not without a few lucrative government contracts, I'd wager - hardly attainable for a hard-working AI services founder.

But here's the thing: the cynicism, while understandable, misidentifies the problem. The question isn't whether early-stage AI companies should copy Palantir's org chart. It's whether there's a structurally new class of worker emerging - one that AI tools are actively creating - and what the right version of that role looks like for the other 99% of the market.

To stress-test the skeptic's view, it helps to understand what Forward Deployed Engineers at Palantir actually do – not the mythology, but the operational reality.

Zoe Scaman's primer on the Palantir model and Nabeel S. Qureshi's first-person essay (both worth reading in full) surface a consistent picture. FDEs embed inside client organizations for months or years. They ignore the stated brief on arrival, operating on the assumption that the client is probably wrong about what's actually broken. They watch people work and infer "revealed preferences" rather than accepting requirements documents. They build fast - usable software within a week or two - to prove they're real before the organization's immune system rejects them.

There are a few other traits worth naming:

Since those essays were written, Palantir has added a "Deployment Strategist": a more traditional consultant with a tech hobby rather than an extroverted engineer with empathy. Sierra calls their version "Software Engineer, Agent." Lovable recently posted for their first FDE. The role is proliferating.

If you're running a 20- or 50-person AI services firm, the FDE as Palantir defines it creates a genuine problem.

Palantir works with Airbus. Their clients have vast, siloed data sets, problems where a mistake can ground a fleet, and the patience - and budget - for months of embedded diagnosis before expecting a working product. The brand and the trust do the pre-selling. Someone else handles the scoping and the commercial conversation. The FDE arrives with a cleared runway and focuses on one thing: solving.

Mid-market clients don't grant that luxury. Their data is less vast, their problems less catastrophic. But the internal politics are often on a surprisingly similar scale. And critically, the commercial dynamic is entirely different.

At an early-stage AI firm, the person walking into a client's office may be walking in before the deal is won – at the proof-of-concept stage, or even earlier. They need to read the room quickly: who actually makes the decision, and how will it be made? They need to run deeper discovery simultaneously with delivery, surfacing problems the client hasn't named yet. They need to own the process from the first workshop through to a working product and employee training on the other side.

And they need to keep a PnL in the back of their mind throughout – identifying upsell paths, managing scope, thinking about what a healthy engagement looks like commercially. That's a skill set that, in a traditional services firm, you'd only find in a very senior consulting partner.

So what does the right version of this role look like for a 50-person AI firm? We call them Forward Deployed Empaths: like an engineer with a real-world MBA.

The key distinction from the Palantir FDE is this: the Forward Deployed Empath needs to be fluent in business problems and comfortable talking PnL and impact directly with decision-makers. That will likely come at some cost to technical depth – but it's a trade-off worth making. In all but the most technically complex cases, a combination of AI tools and support from a more technical colleague will cover the gap.

The best candidates probably aren't coming from Palantir. They look more like the startup generalist: a bit of engineer, a bit of product person, some operations instinct, a genuine discomfort with over-specializing. That profile – which is an asset in an early company and eventually becomes a liability as a firm matures – maps almost exactly onto what an AI enablement practice needs at the client-facing edge.

Two other traits matter. First, unlike traditional FDEs, they can't rely on brand and trust being pre-established. Speed to credibility is the game. Second, their business problem fluency needs to extend to emotional intelligence – reading whether a room is skeptical, politically constrained, or simply confused about what it's actually trying to buy.

The right analogy here may be historical rather than organizational.

Roughly 30 years ago, spreadsheets transformed data analysis from specialist territory into something any business professional could do. VisiCalc and then Excel made complex modeling visual, interactive, and manipulable in real time. The result wasn't just better analysis – it was a structural shift in who held decision-making power. Technical experts were partially displaced by analytical generalists. The modern MBA, in its most useful form, is the professional embodiment of that shift.

I suspect a similar structural shift is underway now. As AI tools lower the technical floor, the determining skill set shifts toward problem diagnosis and emotional intelligence. The engineering barrier that once gatekept this kind of work is eroding. What remains – and compounds – is the ability to walk into a messy organization, understand it quickly, earn trust without triggering defensiveness, and scope a path forward that someone will actually fund.

That's not the same definition of FDE that Palantir popularized. But the majority of companies don't have Palantir's budget or brand to recruit exceptional engineers with exceptional people skills. And, I'd argue, the majority of problems facing organizations adopting AI right now don't require exceptional engineering. They require empathy, commercial acuity, and speed.

FDEs are not hype. The role is real, and it's growing across AI-native product companies and consulting firms alike. What is hype is the idea that copying Palantir's version of the role will produce Palantir's results.

The context is different. The client is different. The commercial dynamic is entirely different. And so the right version of the role is different too - closer to a senior consulting partner's instincts wrapped in a generalist's adaptability, equipped with AI tools that compress the technical gap.

The firms that figure out how to hire, develop, and deploy Forward Deployed Empaths at scale will have something close to a structural advantage in the mid-market. The ones that stick to the old org chart or try to import the Palantir playbook wholesale will find it doesn't travel as well as the brand does.