What You Are Buying When You Hire an FDE
"Forward Deployed Engineer" is on every job board, and the question of whether to hire them has reached the most senior levels of large organizations. Ask three companies what the role is and you will get three answers. For an executive, the definition matters less than a plainer question. What are you actually buying?
An FDE is the distance between what your AI can do and what your customer needs, paid for by the hour.
The last mile is a person
AI products arrive almost working. The model is capable and the platform is general, but every customer's world is specific: their data, their systems, their exceptions, their idea of a wrong answer. Someone has to close that gap.
What your customer needs
That someone is the forward deployed engineer. They sit with the customer, connect the data, adjust the system, and check the answers against what the business actually means. Where the product stops short, they carry it the rest of the way.
That work is useful, and early on it is often essential. It is also a cost, and the cost tells you something.
FDE spend measures the gap
Every hour an FDE spends with a customer is an hour the product could not manage alone. Added up, FDE spend is a measure of how far your AI falls short of your customers.
That is not a criticism. Every new technology needs a person for the last mile. What matters is which way the gap moves over time.
The one test
Does each customer need less FDE time than the one before?
If yes, what your FDEs learn is flowing back into the product. The tenth customer starts from the connectors, templates, and checks the first nine paid for. You are building a product with a services edge.
If no, FDE headcount grows with your customer count. Each deal is a fresh project. You are running a consultancy with a model inside it, and in time your margins will say so.
A product with a services edge
Each customer needs less. The product is learning.
A consultancy with a model inside
Every customer starts from scratch.
I did this job before it had the title
As a developer advocate at Hedera, a distributed ledger technology company, my work was carrying a new technology the last mile into teams that had to make it work. A distributed ledger then and AI now look very different. The last mile is the same shape.
The work that mattered most was never the help I gave one team. It was turning what teams kept needing into things nobody had to ask for again.
| What we built | What it saved the next team |
|---|---|
| SDKs | Writing their own code to talk to the network |
| Demo apps for micro-transactions | Working out the use case from nothing |
| A browser extension for wallets | Building wallet support themselves |
| Integrations with exchanges | Connecting to each exchange on their own |
Each one began as a problem in front of one team and ended as something every team could use. That is the whole discipline: forward deployed first, then folded back.
FDEs extend the work of AI
There is a second reason FDEs matter, and it is the one most executives miss.
FDEs do the work the model cannot yet do reliably. They judge the domain, connect messy data, and catch wrong answers before the customer does. In doing that, they produce the most valuable thing in the system: real examples of what a right and a wrong answer look like for this customer.
Those examples become tests, evaluations, and training data. A model can be swapped next quarter. A set of real cases showing what correct means in your customer's business cannot be bought. It can only be built, and FDEs are the people building it.
So an FDE is not just a patch over the gaps in your AI. Done well, it is how your AI learns each customer, and what they leave behind outlasts the deployment.
Measure what gets folded back
If that is the value, it changes what you measure. Not tickets closed or hours billed. Measure what FDE work turned into product this quarter: the connector that is now built in, the evaluation set that now runs on every release, the template the next customer will start from.
Not this
- Tickets closed
- Hours billed
- Customer satisfaction alone
This
- Connectors now built into the product
- Evaluation sets running on every release
- Templates the next customer starts from
An FDE team that folds nothing back is a services team, however good it is. An FDE team that folds back steadily is shrinking its own job, which is exactly what you want from it.
Hire FDEs. Most AI products need them for now. Just know what you are buying: a person standing in the gap, whose real job is to make the gap smaller. The best sign they are doing it is a question your next customer never has to ask.