Forward deployed engineer vs consultant vs solutions engineer: who gets AI into production

What they hand you at the end. An FDE hands you software running in production inside your tools, while a consultant hands you a recommendation and the build becomes someone else's job.

The table adds two more roles people confuse with FDEs: the vendor's solutions engineer and the product software engineer.

An FDE embeds with one customer and builds inside that customer's systems until a workflow runs in production. Palantir popularized the title for engineers it placed at customers to customize its platform for each one, and the name comes from military language.

The FDE sits with the people who do the work and finds where the real process differs from the documented one, then decides where AI belongs and where plain code or a human approval does the job better.

Before go-live, the new workflow runs in shadow mode next to the old process, and the engagement ends with a handover your team can use. My explainer on forward deployed engineers goes deeper.

A consultant tells you what to do. That's worth paying for when you don't yet know what to build, or whether AI belongs in the process at all.

The trouble starts when the plan meets the system of record. In RSM's July 2026 survey of 1,030 middle-market companies in the US and Canada, 52% already used outside consultants, and firms with moderate or limited pilot success cited data quality (53%) and integration (47%) most. A slide deck doesn't fix either one.

Some consultants do build, so ask what they're measured on and who ends up holding the code. BCG's January 2026 study of 100 private equity investors found that 70% of the successful firms use specialized digital boutiques, while only 45% make sure the knowledge transfers to their own teams. If you're still deciding what to build, my build vs buy guide is a better first step than hiring anyone.

The solutions engineer while you're choosing a product, and the FDE once you've chosen. A solutions engineer works for the vendor before the sale, running demos and proofs of concept, and moves on when the contract is signed.

Use them. They know the product inside out and they cost you nothing. Korn Ferry reported in September 2026 that mid-market PE funds expect AI proofs of concept in 2-4 weeks, and a fast proof of concept on clean demo data shows the model can do the task. Making it work with your data and your systems is a separate project.

Sometimes it is. The integrator SHI and the vendor Cisco told The Register in October 2026 that embedding engineers at customers is nothing new, and DoIT moved its long-running customer reliability engineering practice under the FDE name in late 2025.

Gartner has a name for the relabeling, "FDE washing", meaning ordinary consulting sold under the new title. The test is simple: does the engineer write production code inside your systems, and can your team run it after they leave?

Implementation consultants sit closest to the FDE. They configure and integrate a specific product, which is exactly right when the product already does what you need. When it doesn't, you want someone who starts from your process and not from the vendor's feature list.

Who the work is for. A product software engineer builds features for every customer of one product, and an FDE builds what one customer needs, inside that customer's systems.

The FDE job is also more exposed. OpenAI's FDE postings mention up to 50% travel. Palantir expects its FDEs to spend around 25% of their time on site, according to The Pragmatic Engineer. One recruiter quoted in The Pragmatic Engineer says maybe 10% of engineers want the role, and Paraform puts FDE searches at 8-12 weeks.

If you're an engineer weighing the two, FDE work pays well, and the FDE salary data backs that up. You'll also spend more time with ops people than in your editor, which some engineers love and most don't.

Match the role to where you're stuck.

You don't know what to build yet. Hire a consultant, or take the free AI Reality Check before you spend anything.

You're choosing between vendors. Use their solutions engineers. It costs nothing, and you'll see where each product struggles with your data.

Your pilot works in a demo and stalls on real systems. Hire an FDE.

You're building a product many customers will use. Hire software engineers. An FDE in that seat builds one-offs.

You need a team for a long rollout. Talk to a studio or an integrator, and ask how the handover works before you ask the price.

Ask what they last put into production and who runs it today. Someone who has done the job answers with a system and a name; someone who hasn't answers with a methodology.

Gartner's advice to buyers covers the same ground: settle scope, incentives, governance, ownership, knowledge transfer and the exit plan from day one. The longer checklist is in my guide to hiring a forward deployed engineer.

I'm based in New York, and I don't sell a platform, so the code goes in your repository and the engagement ends with your team running the system.

As an interim forward-deployed engineer I join your team 2-3 days a week for $16,000-$20,000 a month, on site in New York or remote, with the days and the first 30-day goal agreed up front. Scoped pieces like a stalled integration or a set of evals are $175-$250 an hour, and the cost breakdown shows how that compares with a hire. I don't take full-time W-2 roles.

When the work needs more than one engineer, I bring Digital Kitchen, the studio I co-founded, so a bigger rollout still has one senior engineer who knows your process.

I've built 10+ AI products end to end, and at Chatterbox I took over as lead engineer and rebuilt the platform through a seed round and the B2C to B2B pivot. If you're in the city, I can work on site with your team in New York.