TL;DR: Ramp a second forward deployed engineer by making account context self-serve and reserving the senior's time for judgment questions. New engineers take 3 to 9 months to fully ramp on average, and a live account can't wait that long. The playbook below runs a day-one account brief, week-one paired calls, and ownership checkpoints at 30, 60, and 90 days. Spendesk runs its team this way: an engineer picks up any account and knows in five minutes what was promised and what's at risk.
Monthly job listings for forward deployed engineers grew more than 800% between January and September 2025. The candidate pool grew only about 50% in the same window, and filling an open FDE role takes around three months. So by the time your second FDE signs, the first one has been running every account alone for a quarter or more. The accounts didn't pause while you hired.
That's the specific problem this playbook addresses. Not generic new-hire onboarding, but putting a second engineer onto a live account that one person built, where the commitments, decisions, and open risks live mostly in that person's head. Across large companies, per a 2018 study, 42 percent of institutional knowledge is unique to the individual who holds it. On a one-FDE team, the number is closer to all of it.
The whole playbook rests on one distinction. Context transfers, judgment doesn't. Get that split right and the new FDE is useful within weeks. Get it wrong and you burn months of senior time transferring facts a system should have held, while the judgment coaching that actually needs a human never happens.
Why does FDE ramp take so long?
The benchmarks are sobering before you even add the FDE part. A study across more than 80 companies found new engineers take 3 to 9 months on average to fully ramp, and some companies report a full year. The broader knowledge-worker research says the same thing from another angle: new hires get formal training for a couple of months but still take up to six months to fully ramp, with much of that time spent independently hunting for information nobody thought to hand over.
FDE ramp is heavier than either number suggests. A forward deployed engineer is an engineer who spends roughly half their time with customers understanding the problem and half building the solution for that specific customer. That means the new hire is ramping on two things at once, and only one of them is documented.
The two ramps, and which one you control
The codebase ramp is a known quantity. There's a repo, a README, tests, and a senior engineer who can review pull requests. Every engineering org has run this before.
The account ramp is where the shared context problem bites. What was promised during the sales cycle, what was decided in week three, which integration got descoped and why, which stakeholder is nervous. None of that lives in the repo. Most of it isn't written anywhere, and what is written sits in a tracker last touched before kickoff. This is the same gap that caps how many accounts each FDE can carry, and it's the part of ramp you can actually compress.
Context transfers, judgment doesn't
An FDE at a vertical AI startup put the hard part plainly when we asked about scaling his team: "being given the same context does not yield the same sort of, like, a judgment call." His open question was "how I can quickly bring the FD up to speed," and the honest answer has two halves.
The half a system can carry
Context is the transferable half. What was promised, what was decided, what's at risk, what happens next, each with the source behind it. That's account state, and shared context exists precisely so a new person can self-serve it instead of extracting it interview by interview.
The half that needs a human
Judgment is the other half, and it doesn't move by reading. Knowing which commitment to renegotiate, when to push back on a scope request, how hard to press a champion who's gone quiet. That transfers through reps: paired calls, debriefs, and small decisions made with a senior nearby.
The playbook's job is to stop spending senior time on the half that does transfer. Every status question the new hire brings to the senior costs more than the answer, because interrupted work takes an average of 23 minutes and 15 seconds to resume. An FDE team lead at a roughly 50-person AI company told us his success metric for tooling was exactly this: "the ability or ease of bringing on a second FTE and getting them ramped up to either replace or augment the existing project."
What the new FDE needs before their first customer call
Before the first call, the new FDE needs a floor to stand on: an account brief that answers four questions with sources: what was promised, what was decided and why, what's at risk right now, and what happens next. If they can answer those without asking anyone, the first paired call becomes coaching instead of narration.
Where the brief comes from matters as much as what's in it. A brief the senior writes by hand is a snapshot that starts decaying immediately, and writing it is exactly the admin work you hired to reduce. An account brain built for FDE teams assembles the brief from the event stream instead, from calls, Slack, and tickets, with each fact linked to its source. Where Tandem agents draft any part of it, a human approves before it guides customer-facing work.
This is what Spendesk describes on their team: "An engineer picks up any account and knows in five minutes what was promised and what's at risk. Nobody maintains anything," per Mathilde Hamy, Head of Customer Success and Account Management at Spendesk. Their call prep dropped from 45 minutes to 5 on the same footing.
The ramp playbook, day one to day ninety
Here's the sequence, built so senior time goes to judgment and everything else is self-serve.
- Give the account brief before the office tour. On day one, the new FDE reads the live account state themselves, commitments, decisions, risks, next steps, and sources, without a meeting on the calendar. The senior's first session then starts from "what surprised you?" instead of a two-hour history lecture.
- Pair on every call in week one, with the brief as the shared floor. The new FDE preps from the brief and drafts the agenda. The senior corrects what the brief can't hold, the relationship texture and the unwritten history. Prep from a maintained brief takes minutes, not the better part of an hour.
- Route status questions to the brain, judgment questions to the senior. Make the rule explicit on day one. "What did we promise about SSO?" goes to the system. "Should we renegotiate the SSO date?" goes to the senior. This single rule protects the senior's build time from the 23-minute interruption tax.
- Hand over one owned workstream in week two. Small, reversible, and on the live account, a data-mapping task or one integration. Real ownership forces the new FDE to exercise the context, which is how you find the gaps in it.
- Run a 30-day checkpoint where the new FDE walks the account back to you. They narrate the state: what's promised, what's risky, what they'd do next. This is a readiness check for their benefit, not a reporting ritual. Gaps in the narration tell you what to pair on next.
- Transfer first-response ownership by day 60. The new FDE takes customer questions first and pulls the senior in by exception. The senior shifts to reviewing decisions rather than making them.
- At day 90, decide: keep ownership or graduate the account. The head of an FDE team at a finance-software company described running exactly this model, where an FDE owns an account for its first 90 days and then graduates it to a shared pool. Whether you run a pool model or dedicated ownership, day 90 is when the account should survive a handoff to anyone on the team.
Where does ramp context come from today?
Most teams run ramp on some mix of the options below. The table is honest about what each costs.
- Spreadsheets + Slack. How the new FDE learns the account: Scrolls channels, asks the senior. Senior FDE time cost: High and interrupt-driven. Freshness: Stale between manual updates, varies by team.
- Shadowing every call. How the new FDE learns the account: Osmosis over weeks. Senior FDE time cost: Highest, doubles the call load. Freshness: Fresh but nothing gets recorded.
- Handoff doc or wiki. How the new FDE learns the account: Reads a snapshot. Senior FDE time cost: Hours to write, upkeep unquantified. Freshness: Decays from the day it's written.
- Account brain, queried from anywhere. How the new FDE learns the account: Asks a question, gets a source-linked answer. Senior FDE time cost: Judgment questions only. Freshness: Synced from calls, Slack, and tickets.
The pattern behind the first three rows shows up in the enterprise AI data too: generic tools stall because they don't learn from or adapt to the team's workflows. A wiki doesn't know a call happened. A spreadsheet doesn't know the scope changed.
The fourth row is what Aircall reports after moving rollout context into a queryable team memory: "We're seeing engineers confidently run rollouts they've never been on a call for. The memory belongs to the team, not the person," per Tom Chen, CPO at Aircall.
What stays manual
Honesty about the limits keeps the playbook credible.
Judgment stays manual. No brief teaches a new FDE when to hold a date and when to trade it. That comes from paired calls and debriefs, which is exactly where the reclaimed senior time should go.
Relationship trust stays manual. The customer trusted a person, not a team. The new FDE earns that on calls, and the first hard conversation (a slipped date, a descoped feature) should still be paired.
Curation stays manual. Someone decides what's worth keeping in the brain, and any update drafted by an agent gets human approval before anyone relies on it. The 30, 60, and 90-day checkpoints are conversations too. A system can prepare them, it can't have them.
Ramp the second hire without burning the first
The benchmarks say months. The compressible part is the context half, and compressing it is a systems decision, not a training decision. Give the new FDE self-serve, source-linked account state on day one, pair on judgment, and check ownership at 30, 60, and 90 days. Codify what worked into your FDE operating system so the third hire ramps faster than the second.
See how fast a new FDE could pick up one of your live accounts: book a demo and watch Tandem build the day-one brief, risks and sources included, from your real account history.
FAQs
How long does it take to onboard a forward deployed engineer?
Full ramp for engineers averages 3 to 9 months, and FDEs carry an extra account ramp on top. The account half compresses to weeks when context is self-serve and source-linked. The judgment half still takes months of reps, but the account keeps moving while it builds.
Should the new FDE shadow every customer call?
Shadow every call in week one, then stop. Shadowing transfers texture but doubles the call load and records nothing. After week one, the brief carries the state and pairing continues only on high-stakes calls.
What if the first FDE has no time to document anything?
Then don't ask them to. A written handoff pack is the first thing to slip and the first thing to go stale. Sync context from what already happens (calls, Slack, tickets, working sessions) so the senior's only job is answering judgment questions.
Can a second FDE really run an account they've never been on a call for?
Aircall reports exactly that: engineers confidently running rollouts they've never been on a call for, with 3x rollouts per engineer as the outcome. It works because the state is queryable with sources, so the engineer can verify rather than trust. The first customer-facing call still deserves a paired warm-up.
Does self-serve context replace ramp conversations?
No, it replaces the narration portion of them. The senior stops reciting history and starts coaching decisions. The conversations get fewer and better, they don't disappear.
Key terms
Forward deployed engineer (FDE): An engineer who spends roughly half their time with customers understanding the problem and half building the solution for that specific customer.
Ramp time: The period between a new hire's start date and full productivity. For engineers it averages 3 to 9 months, and for FDEs it includes a second ramp on account context.
Account brief: A short, source-linked summary of a live account answering what was promised, what was decided, what's at risk, and what's next. Useful for ramp only if it's generated from live account state rather than written by hand.
Shared context: The account's memory held at team level, so any person or agent can query what was promised, decided, and at risk. The alternative is context held by one person, which is what makes second-FDE ramp slow.
90-day ownership: A team structure where an FDE owns a new account for its first 90 days, then graduates it to a shared pool. It only works if the account state survives the graduation.
Project brain: The category term for a maintained, queryable store of project facts with sources. An account brain is Tandem's version, connected to calls, Slack, and tickets.