Harvey's Forward-Deployed Model: Lawyers, Not Just Engineers
Legal AI vendor Harvey (Series C, $1.5B valuation, 1.3K employees) is staffing every deployment with former practicing attorneys. Chief Product Officer Anique Drumright confirmed roughly 180 legal engineers on the team, most with 8-10 years of practice experience.
The comp is public. Harvey posts $220k-$320k OTE on a 75/25 split, plus equity. Minimum bar: JD plus 3 years at a top-tier firm or in-house.
What they actually do. The function splits three ways: pre-sales discovery, post-sales implementation, and custom solution pods. For complex deployments, Harvey adds mixed pods with a product manager, 1-2 lawyers, and engineers working bespoke for the account.
Every customer gets a legal engineer, not just enterprise accounts. With 1,400+ customers and 100,000+ lawyers on platform (60%+ of AmLaw 100), that is a structural choice, not a pilot programme.
Why it matters for sales roles. This is not a customer success engineer play or a solutions architect with domain training. Harvey is hiring people who ran legal matters for a decade, then converting them into a hybrid pre/post-sales function. The credibility gap in complex enterprise sales typically takes 6-9 months to close. Harvey skips the ramp.
Drumright's framing: the legal engineer surfaces workflows a software person would not know to ask about, and adoption barriers a customer would not volunteer to a vendor.
The part nobody noticed. Harvey is now certifying legal engineers who do not work at Harvey. That extends the deployment model beyond headcount.
Context for ANZ. Forward-deployed engineer models (Palantir's original playbook, now spreading across enterprise AI) are showing up in sales comp structures as companies figure out agents do not deploy themselves. If you are evaluating roles in AI sales, watch for:
- Deployment capacity as a quota constraint
- Post-sale implementation flagged in the sales cycle
- Comp structures that blend sales and services (like Harvey's 75/25)
- Domain expert requirements showing up in SE and CSM job descriptions
Harvey's model costs more upfront but removes the technical/domain translation layer in enterprise deals. Worth watching whether that comp floor ($220k base equivalent) becomes the benchmark for forward-deployed roles in vertical AI.