OpenAI Codex, powered by frontier models, enabled Asana to remove its legacy Enzyme testing system in roughly two weeks — a migration previously estimated to take five years and cost about $6 million. Asana, a work-coordination platform that uses AI agents and automations for customers, applied the same approach internally. The final model and infrastructure cost was approximately $12,000.

From a five-sentence prompt, up to four coding agents worked in parallel, each in a separate copy of the codebase. An engineer checked progress twice a day and reviewed every proposed change. Simpler instructions worked better than a more elaborate setup. After 1.5 weeks of engineering effort spread across two calendar weeks, Enzyme was fully removed.

The experience changed which long-running software projects Asana considers practical to take on. The team plans to test agents on other migrations, rewrites, and performance problems it once assumed would take years, while keeping human review in place. Asana CTO Amritansh Raghav cautioned that not every years-long project will collapse into weeks, but agents can give engineers more room for craft and make once-impossible work worth attempting.