What is forward deployed engineering?

Forward deployed engineering is an approach where software engineers work inside a customer's business instead of building from a distance. A forward deployed engineer (FDE) spends time with the people doing the work, learns how processes really run, and then builds and deploys software — increasingly including AI — directly into those workflows, improving it based on real use.

Where forward deployed engineering came from

The idea was pioneered by Palantir, which sent engineers to work directly with customers rather than handing over software and walking away. Those engineers learned each customer's problems first-hand and adapted the product to fit — work that was hard to do from a head office.

More recently, forward deployed engineering has become central to how AI companies — including OpenAI and Anthropic — help organisations adopt their technology. Powerful AI models are only useful once someone connects them to real data, real processes and real people. That last mile is exactly what forward deployed engineers do.

What does a forward deployed engineer actually do?

A forward deployed engineer's work usually looks like this:

• Learning the business: sitting with operations, finance or support teams to see how work really flows — including the workarounds nobody wrote down.
• Finding the problem worth solving: spotting repetitive tasks, delays and errors, and working out what each one costs.
• Building quickly: creating a working solution — an automation, an internal tool, an integration, an AI assistant — rather than a report or slide deck.
• Deploying it for real: putting it in front of the people who'll use it, training them, and fixing what doesn't work.
• Measuring and iterating: checking whether it actually saved time or reduced errors, then moving to the next problem.

FDE vs software engineer vs consultant

Forward deployed engineers are sometimes compared to solutions engineers too. The difference is that solutions engineers mostly help sell and configure a product, while forward deployed engineers build whatever the customer's problem needs.

Why forward deployed engineering works

• The most important knowledge isn't written down. It lives in how a dispatcher allocates jobs or why finance double-checks certain invoices. You only learn it by being there.
• Short feedback loops. Engineers see immediately when something doesn't work and fix it — instead of discovering problems months later.
• Higher adoption. Software built with the people who'll use it fits how they work, so they actually use it.
• Clear accountability. The engineer who built it is still there, measured on results.

When should a business use forward deployed engineers?

Forward deployed engineering is a good fit when:

• You know something is slow or costly, but haven't defined exactly what software would fix it.
• Your operations run on spreadsheets, email and chat groups.
• Previous software projects delivered something your team didn't use.
• You want to use AI but don't know where it would genuinely help.
• You don't have an in-house engineering team, or yours is busy with the core product.

It's less necessary when requirements are already precise and stable — for example, rebuilding a well-documented system on new technology. In that case, a traditional project team can work well.

What it looks like in practice

A typical forward deployed engagement follows five steps, repeated until the business runs smoothly:

1. Understand how the business really runs, end to end.
2. Find the repetitive work and rank it by time lost.
3. Automate or build the fix, with checks for anything important.
4. Deliver it to the team, with training and quick adjustments.
5. Measure what changed, then move to the next task.

For example, a team that receives orders by email and WhatsApp might have someone retyping each order, checking stock in a spreadsheet and confirming by hand. A forward deployed engineer would see this first-hand, then automate it so orders are read and entered automatically, stock is checked instantly and customers are confirmed in seconds — with people only stepping in for exceptions. We walk through this in detail on our What we solve page.

Forward deployed engineering for growing businesses

Forward deployed engineering began with large enterprises and well-funded AI companies, but the same approach works — arguably even better — for growing businesses. Smaller organisations have less documentation, more knowledge in people's heads, and less tolerance for software that doesn't fit. A small, experienced team working inside the business can make a big difference quickly.

Skills a forward deployed engineer needs

Forward deployed engineering asks for an unusual mix of skills. Strong FDEs combine:

• Broad engineering ability — building across front end, back end, data and integrations, often in the same week.
• Practical AI experience — knowing where AI is reliable, and how to add checks where it isn't.
• Business understanding — reading a process and seeing where the time and money go.
• Communication — explaining options in plain language to people who aren't technical.
• Judgement and ownership — deciding what not to build, and caring whether the result is actually used.

Common misconceptions

• "It's just consulting with a new name." Consultants hand over recommendations; forward deployed engineers hand over working software.
• "It's only for huge enterprises." Smaller businesses often benefit more, because so much of how they work lives in people's heads.
• "It means engineers sitting in your office every day." Being close to the work matters more than the location. Much of it can be done remotely, with time on-site when it helps.
• "It's slower because there's no big upfront specification." In practice it's often faster, because less time is spent building the wrong thing.

How forward deployed engineering is priced

Pricing varies by provider. Common models include a fixed price for a clearly defined first project, a monthly fee for an ongoing embedded team, or a project that ends with handover to your own staff. Whatever the model, look for clear goals agreed upfront and regular evidence of progress — working software, not just reports. If you're weighing up an external team, our guide on hiring a software development company in India covers contracts, ownership and red flags.

How to evaluate a forward deployed engineering partner

If you're considering working with an external team, ask:

• How will your engineers learn how our business actually works?
• Who will we work with day to day — and are they the people who'll build it?
• How do you decide what to build first, and how do you measure success?
• How will our team be trained, and what happens after launch?
• Who owns the code and data, and how is access to our systems controlled?

For a broader checklist, see our guide on how to choose an automation partner, or learn more about how we work as forward deployed engineers.