Decagon Challenges Big Tech’s Forward-Deployed AI Push

Forward-deployed engineers, or FDEs, combine software development with consulting and product skills, embedding with customers to connect AI models to proprietary data, workflows and security controls. The role has become a focal point in enterprise AI because many companies can run promising pilots but struggle to move them into production. The hiring boom also frames a deeper dispute: whether deployment failures reflect immature technology that vendors must simplify, or customers’ lack of organizational readiness and technical capacity.

AWS on June 30, 2026, committed $1 billion to a Forward Deployed Engineering organization with thousands of specialists. Microsoft followed on July 2 with the $2.5 billion Frontier Company, comprising about 6,000 engineers and experts. Decagon, a customer-service AI startup, challenged the premise on June 22, saying bespoke deployments had become a scaling bottleneck. Its Agent Development team shifted most build work to Agent Development Managers and cut custom engineering per AI agent by 80%, suggesting embedded engineers should make clients self-sufficient rather than remain a permanent layer.