Perspective

The Forward Deployed Engineer Is Half the Team

Most vendors hire the engineer who builds and skip the strategist who readies the customer to change. Why forward deployment needs a team of two.

Harsh AgarwalNikhil Sarma
Harsh Agarwal & Nikhil Sarma
Co-authors
October 16, 202612 min read
Black and white photograph inside a rally car, the driver steering while the co-driver reads pace notes and calls the road ahead.

The co-driver calls the road before the driver can see it.

01. Context

The change that is upon us

For two decades, Enterprise SaaS was sold on one promise: configure, do not customize. A standard product, set up well, would be enough. Much of the Solutions Engineer’s job was to demonstrate that before signature and pricing models assumed it would hold after.

That promise is breaking. Vendors are hiring Forward Deployed Engineers (FDEs), placing them at customer sites for weeks at a time, and committing to build what the product does not yet do. Palantir, which has run this model for twenty years, is now cited as the blueprint.

Some of this reflects real need, because enterprises keep producing requirements nobody can specify in advance. Some of it is fear of being left behind.

Either way, a vendor that hires an FDE is admitting that configuration was not enough, and that admission has a price.

The questions organizations now face

When a vendor commits to building inside a customer’s environment, several questions arise at once. Most companies are answering them by trial and error.

  • Is forward deployment a job for one person or two? Who owns the technical build, and who owns whether the organization actually changes?
  • If you build a forward-deployed team, what happens to your Solutions Engineers?
  • Which deals justify the cost, and which customers will waste it?
  • How should a CRO or VP hire for these roles, structure the team around them, and know whether the change is working?
  • Who pays for the build, when the pricing model was designed to assume construction?

What this article tackles

We do not have complete answers, and we are suspicious of anyone who claims to. We write from a specific vantage point: enterprise software sold into large organizations attempting serious transformation, not incremental change. Other contexts exist, and other models may well work there.

We have split the piece into two voices on purpose, so you know whose view you are reading. Harsh writes from within the SE org. He covers the two-role model, the paths open to Solutions Engineers, and which deals can carry the cost.

Nikhil takes the Revenue Organization's perspective. He covers how to qualify the customer, who owns what, how to structure and hire for the new roles, and how to measure the result.

Pricing deserves more room than we can give it here. We set out the problem at the end and will address it in a follow-up article.


02. The Two-Role Model and Where It Applies

Author: Harsh Agarwal

I am writing this from inside the change. This year my own role moved away from classic solutions engineering. Junior presales engineers on my team now build working prototypes with prospects in two-week workshops, using AI coding tools. And I have been asked to stay on the accounts I sell through the blueprinting phase. The message is simple: “you painted this picture, now help make it real”.

Palantir sent two people, not one

Palantir rejected the configure-don't-customize model on purpose. Around 2006 its defence and intelligence customers could not say what they needed, could not share their data, and changed their workflows constantly. So they started shipping engineers in addition to the software. For about a decade it had more Forward Deployed Software Engineers, internally called Deltas, than product engineers.

Enterprise AI recreates those conditions on almost every large deal today: requirements nobody can fully specify, and a platform that can do almost anything. That is why a twenty-year-old model is being copied.

Here is the detail most people copying it miss. Palantir deployed a team of two.

Deltas worked along with Deployment Strategists, called Echos, who owned adoption, internal politics and executive alignment.

A Delta alone builds something technically correct and operationally irrelevant. An Echo alone strives for alignment with nothing actually shipped.

Most of the FDE conversation is about the Delta: send an engineer, build something impressive in a couple of weeks. That makes an excellent demo of what is possible, but does not account for the iteration and change management necessary for adoption. That is the Echo’s job, and I think it is the harder and more valuable one. Most companies are trying to do both with one hire.

Two paths out of the SE role

Two forces are squeezing the SE role. First, proving technical fit has become cheap. When a working prototype takes an afternoon instead of three weeks, the demo is no longer scarce. Second, buyers who have lived through a failed deployment no longer ask whether the software can work. They ask who will make it work, and the answer is not the person who hands over the account at signature.

So if you are an SE, my honest view is that change is coming and you should choose a path. The Delta and Echo pairing gives you two.

The Delta path is for the SE who loves to build. This is less about coding than it used to be, because AI tools now do most of the heavy lifting. What matters is a builder's mindset: knowing exactly what you want, noticing when it does not work, and asking whether there is a better way. You can learn the tools in weeks. The harder part is the habit of shipping rather than presenting.

The Echo path is where I think most SEs belong, and it is the one I have taken. Compare the Echo's job with what a good SE already does: mapping stakeholders, building the business case, working a buying committee, turning technical capability into business consequence. It is much closer to SE strengths than production engineering is. There is no course for it. You build domain depth, the ability to read a room, and the willingness to stay in that room after signature.

This path has a cost. An SE who stays through deployment runs far fewer deals before signature, so presales headcount starts to grow with closed deals rather than pipeline. Splitting the role into a pre-sale SE and a post-sale SE looks like a fix, but it only creates another handover.

Who this is for

Whenever I read that something "is the way", I want to know: for what product, and at what deal size? Here is my answer.

This model makes sense only for large enterprise deals, at least USD 300K in annual recurring revenue. The reason is margin. FDEs are expensive engineers who sit on a customer site for weeks at a time. Add a strategist who stays on the account for months, and the cost of winning one customer rises sharply. The contract has to absorb that.

In mid-market deals, the numbers do not add up. The model stretches a sales cycle that should be short and pushes acquisition cost past any sensible return. The buyer does not want it either. Opening its systems to outside engineers, running the change that follows, and getting a vendor's access to sensitive data past its security team is a heavy price for a contract that small.

One exception to deal size: deploying a team for a limited Proof-of-Concept for a large enterprise can pay off if it opens the door to a much bigger contract. Think of a global company that pilots every new system in one market before rolling it out across a region. Treat it as a bet on the account, not a deal on its own terms.


03. Qualifying, Owning, and Staffing the Work

Author: Nikhil Sarma

Harsh has described this change from inside the SE team. I want to look at it from where the CRO and the VP of Solutions sit: which deals deserve the investment, who owns what, whom to hire, and how to tell whether it is paying off.

Qualify before you deploy

Poor qualification has always been expensive: demos for prospects who are never going to buy, PoCs that go nowhere, SE weeks lost to stalled deals.

Now put a Delta and an Echo on a customer site for four weeks: two senior people, fully loaded, on a single account. If the deal does not close, every one of those weeks is acquisition cost. If it closes but never expands, you have spent the first year’s margin building software for one customer. The model pays only when the customer buys, uses what was built, and grows the account.

So treat the decision to deploy a pair as part of qualification. Before committing, answer three questions.

Do the customer’s senior leaders know why they need this? They should be able to link the change to their CEO’s or board’s priorities: competing better, running the business better, or reducing operational risk.

Does this customer have the capacity to change? Have they done something similar before? What other projects are competing for resources? Who on their side will still own this a year from now?

Is there an expansion path beyond the initial deal? That means a named next market, business unit or phase, and someone with the authority to open it.

If you cannot answer these, you are giving away expensive consulting. As Harsh puts it, “technically correct and operationally irrelevant”.

Who owns what

When a Delta goes in for a paid pilot or a signed rollout, an Echo goes with them. The Delta builds. The Echo makes sure the customer’s organization is ready to use what is being built - the right executives engaged, the people who will run it on board, and someone ready to decide when the build is done. The two jobs run in parallel.

Is that not the AE’s job? Partly.

The Echo goes much deeper, learning how teams work, where the resistance sits, and what each executive needs to see before backing the change. Few AEs have the time. And the customer sees the AE as the seller, which gets in the way of an honest conversation about why a team will not use something.

So the AE owns the commercial relationship and the Echo owns the customer’s readiness to change. On a few strategic accounts one person might do both; elsewhere you need two.

Ownership also shapes the handover after signature. Most companies treat it as a documentation problem, but what gets lost is the decision context. Why did the customer buy? What did the sponsor promise their own leadership? Which teams agreed reluctantly? That knowledge lives in people, not documents. The Echo carries it beyond the signature because they were in the conversations that surfaced it.

I learned this during my years as a Field Application Engineer at Dolby, where customer relationships lasted over a decade. The people who helped the customer make the decision were also the people who supported them through the implementation. I was the Delta and shared the Echo’s job with the AE, who handled the CxO conversations while I worked with the Operational and Management teams who had to make the technology work day to day. The Echo’s role can be shared like that, but only if everyone has clear ownership.

Structuring and hiring for the two roles

Customers rarely lack information. They lack the clarity to decide and the confidence to change.

The Delta creates that clarity by making the outcome tangible; the Echo by making the change workable for the people who have to live with it. Hire people who judge their work by the customer’s decision, not by the quality of what they presented.

Where they sit. Both roles report into Solutions, not Sales. An Echo on a sales comp plan will lean toward closing. Customer Success is the wrong home too, because the Echo must be in the account before signature. The Delta sits alongside the Echo, not in product engineering, so the two are staffed as one unit.

Hiring the Delta. While it is true that coding skills matter less than they used to, screen for judgement when the brief is ambiguous. Give candidates a vague brief: strong ones ask what the business is trying to change before they open a tool. Weak ones start building at once, and build the wrong thing perfectly.

Hiring the Echo. This is the harder hire. The core skill is discovery - finding out what someone will not say aloud. Which executive is quietly opposed? People learn through experience, often by getting it wrong. Ask candidates to walk you through a deployment that stalled and what they missed. Notice whether they describe the customer’s organization, or only their own product.

Where to look. Look within, but also beyond the SE org. For the Delta, product engineers who already gravitate toward customers make strong candidates. For the Echo, look at experienced SEs who own the business case, and at Professional Services consultants who are good at drawing out requirements and building consensus across teams. They have already handled the politics of deployment after signature; the Echo role puts that skill to work while the decision is still being made.

04. Measuring It and What Comes Next

Authors: Harsh Agarwal and Nikhil Sarma

How do you know it is working?

12 months in, the CRO and CFO will ask for evidence that this has worked. Agree on that evidence with Sales and Solutions at the start.

The VP of Solutions should watch the leading indicators:

  • How many workshops and pilots turn into contracts
  • How long it takes from the end of a build to the customer’s decision
  • How long implementation takes compared with deals without a forward-deployed team
  • Whether the customer is still using what was built three months after rollout

The CRO and CFO will watch the lagging indicators:

  • Average deal size and win rate above the deal-size threshold
  • Expansion revenue from forward-deployed accounts, compared with the cost of the Deltas and Echos on them
  • Margin on those accounts at 12 and 24 months

Read the two sets together.

If the VP’s numbers look good and the CFO’s do not, the model works but you are choosing the wrong customers, so go back to the three qualifying questions.

If the VP’s numbers are weak, the problem is execution - the pairing, or the people you hired.

If both sets look good, you have earned a harder conversation.

What comes next

That conversation is about price. Configuration was the promise, and every forward-deployed team you hire is proof it was not enough. Someone has to pay for that.

Bill the build by the day and you reward slow deployments. Fix the fee and you are pricing a scope that will change many times. Fold it into a higher usage rate and you need a minimum commitment underneath. None of these fits cleanly, which is why pricing will get an article of its own.

Most vendors are paying the bill themselves, hidden in acquisition cost where nobody has to explain it.

The three qualifying questions will not tell you how to price the build, but they will tell you whether a customer is worth building for. That call cannot wait for the pricing to be settled.

Ready to apply this?

Book a diagnostic conversation.

Loading booking calendar...