A practice prompt we wrote. No company or candidate report names it, so it carries no company tag.

How to answer

Two Anthropic postings show why the judgment matters. The Pre-Sales Program Lead posting describes the company’s team as having more demand than it can serve. Source 1Pre-Sales Program Lead, Forward Deployed EngineeringPublisherAnthropic (Greenhouse job board)Source typecompany job posting The Munich “Applied AI Engineer, Enterprise” posting says that once a customer decides to build on Claude, the engineer sets up the customer’s systems integration partner and stays on as technical advisor while the partner does the build. Source 2Applied AI Engineer, Enterprise (Munich, Germany)PublisherAnthropic (Greenhouse job board)Source typecompany job posting Neither says how the company decides where engineers go. Together they show that its FDE time is scarce and that handing the build to someone else is a normal path.

  1. The cost of , in the customer’s terms. Their people pulled into working sessions and access reviews, and a system built by someone who leaves. Lead with this. Your team’s capacity is a real second reason, but give it to your own side, such as the account executive, never to the customer.
  2. Three or four conditions, said as tests. The product as configured covers the need, and the gap is setup or training. Their engineers can own it, and what blocks them is an answer, not labor. Many customers share the gap, so it belongs in the product. No one on their side owns the outcome, so an embed would leave with the engineer; that is a reason to find an owner first, not to leave the customer alone.
  3. What you offer instead, even to a customer who will pay. A short working session, office hours, a reference implementation, a partner, a product request with a date, or a time-boxed embed with exit criteria. “No” should come with a smaller, specific offer.
  4. One time you did it. What you saw, what you said and how it turned out.
  5. The reverse signal. What would have made you embed after all.

The trap is “almost never”, which sounds like defending the role. The opposite trap is deal size: judging customers by revenue rather than need.

GlossaryForward deployed engineerA software engineer who builds and ships production systems inside a customer’s problem and environment, accountable to that customer’s outcome.More on Forward deployed engineerGlossaryEmbeddingA vector representation of text or other data used to measure similarity of meaning.More on Embedding

Follow-ups

What the interviewer may ask next, once your first answer is on the table.

  • The account executive wants the embed because it helps the renewal. What do you say to them?
  • How would you know, three months later, that self-serve was the right call?
  • Their CIO insists on an embedded engineer and will pay for one. Do you still say no?

Where answers go wrong

  • Answers “almost never”, which sounds like defending the job rather than judging the customer’s need.
  • Makes it a question of deal size alone and never names what embedding costs the customer or what they get instead.

Answer this in two minutes

Write the answer you would say out loud. The clock starts with your first word.

Two minutes

Illustrative answer about a fictional project

Embedding costs a customer more than our time. It pulls their people into working sessions, puts us through their access reviews and leaves them a system built by someone who will leave.