Augment Code

Augment Code publishes AI coding agents, and it also publishes the manuals for adopting them. Two documents carry its name: "AI-Powered Engineering at Scale: The Adoption Playbook" (November 2025), a four-phase roadmap running from a handful of internal champions to AI treated as invisible infrastructure, and "How AI Changes the SDLC: A Six-Stage Guide" (June 2026), signed by Paula Hingel. Both are pitched at CTOs and VPs of Engineering. The company's agent work is recorded as built on Claude Code, and it sits in the graph as a strategic partner.

It is also cited as evidence by others. Anthropic's 2026 Agentic Coding Trends Report uses Augment Code to illustrate the collapse of onboarding time: a project estimated at four to eight months was delivered in two weeks.

What makes the company's own writing travel is that it names failure modes rather than wins. Hingel's guide leans on the DORA 2025 report, which correlates AI adoption positively with throughput and product performance but negatively with delivery stability, and concludes that process maturity counts for more than the tool. It names three structural risks: erosion of the junior pipeline, governance gaps at scale, and circular validation, where AI-generated tests confirm AI-generated code instead of checking the real requirement. That last term escaped the guide. SFEIR's July 2026 note on code review adopts it as the named failure mode of the constraint ring around agents.

A vendor of coding agents arguing that the bottleneck is verification, not generation, is the position worth sitting with.