Guide

Why AI startups run out of deployment capacity, and what to do about it

6 of the 46 calls were about forward deployed or implementation engineers. The story was the same each time: customers wanted it, and the team couldn't deploy fast enough.

By the Lyra Capital team · · 3 min read

The bottleneck is deployment

AI startups selling to enterprises told us the same story. Pilots get signed, and each one needs custom data work, integrations and workflow tuning. Core engineers get pulled into deployments and the product stalls. That's the job forward deployed engineers exist for. If the term is new, start with what a forward deployed engineer is.

Working workflow first

One founder building a governance layer for enterprise AI noticed that their buyers didn't have working AI workflows yet, so there was nothing to govern. The lesson travels: deploy the workflow that produces value first, then sell the platform on top of it.

Let existing customers set the roadmap

Teams with live customers had a backlog of implementation requests. Those are the clearest specs you will ever get. An engineer working through them produces revenue and product insight at the same time.

How to scope a forward deployed engagement

  1. Pick one customer deployment with a clear success measure: a workflow live, a metric moved.
  2. List the systems and data the engineer needs, and start approvals early. Enterprise security reviews are slow.
  3. Decide what goes back into the core product and what stays custom.
  4. Set a weekly demo with the customer.
  5. Plan the handover: docs and tests, so the next deployment is faster.

Your own team or a partner

If every customer deployment needs custom work for the long run, you need forward deployed engineers of your own. If you need them now, while you hire, a partner is faster. Lyra places forward deployed engineers in startup teams, and they can start in days. Hiring a senior engineer in-house usually takes 4 to 12 months.

Questions founders ask

What does a forward deployed engineer do at an AI startup?

Ships the integrations, data pipelines and workflow logic that make the product work inside a customer's environment, then feeds what they learn back to the core team.

When should an AI startup hire forward deployed engineers?

When signed pilots are waiting on implementation and core engineers keep getting pulled off the product to deploy.

How fast can a forward deployed engineer start?

Through Lyra, in days. Hiring a senior engineer in-house usually takes 4 to 12 months.

Based on Granola notes from 46 calls Nam Dao and the Lyra team held with founders about hiring engineers, 6 July to 7 November 2026. Names, companies and figures are left out. AI helped pull the questions from the notes, and each one was checked against its note. Facts about Lyra engineers are as published on lyratechnologies.com.au.