Ramp Engineers Use Codex with GPT‑5.5 to Accelerate Code Review and Build On‑Call Assistant

Ramp’s engineering team has adopted OpenAI’s Codex powered by GPT‑5.5 to slash code‑review turnaround from hours to minutes, citing the model’s deep reasoning and thoroughness. Codex reviews pull requests automatically, catching issues that human reviewers often miss, and can be accessed via a CLI or a dedicated app with visual cues. The team also leveraged Codex to develop an On‑Call Assistant that handles business logic, concurrency bugs, and incident investigations, dramatically reducing the mental load on engineers during on‑call rotations. Codex’s ability to reason through complex domain knowledge enables rapid prototyping of internal tools, allowing Ramp to iterate faster on product features. The engineers emphasize that Codex’s output is “industry gold standard” and has become a mandatory part of many code‑review flows. By guiding engineers through a solid first session, Ramp built trust and encouraged experimentation with AI tools. The success demonstrates that AI can become a core part of engineering workflows, shifting the role of engineers toward orchestrating AI rather than writing every line of code. Ramp plans to continue expanding Codex’s use to further accelerate development velocity.