The model
Harvey AI — the legal and professional services AI platform — employs FDEs to embed within Am Law firms, in-house legal teams and professional services organisations. FDEs map out legal workflows, de-risk constraints, build custom retrieval pipelines, and integrate with client systems including document management systems, KM platforms and SSO.
Harvey's FDE model runs alongside a separate Legal Engineer track — practising lawyers serving in an advisory capacity — creating a two-track embedded services structure that covers both the technical and domain dimensions of client engagement. Switching costs are high once Harvey is integrated into a firm's document and knowledge stack, which amplifies the value of the embedded engagement.
Strengths & weaknesses
Strengths
- Deep domain specialisation in legal workflows provides genuine differentiation against generalist AI platforms.
- Dual FDE + Legal Engineer structure covers both technical and domain dimensions of client embedding.
- Clients include large Am Law firms where switching costs rise sharply once Harvey is integrated.
- Strong alignment with Harvey's platform-first approach to legal AI.
- Custom retrieval pipelines plus DMS/KM/SSO integration anchor stickiness inside the firm's tech stack.
Weaknesses
- Legal-sector AI adoption is slower and more risk-averse than other verticals.
- FDE work requires specialised knowledge of legal process, DMS and regulatory constraints — limiting hiring speed.
- The combination of frontier AI capability and legal-workflow expertise is a scarce talent pool.
- Two-track structure adds coordination overhead between FDEs and Legal Engineers.
- Vertical focus caps total addressable market versus horizontal AI-platform peers.