Observed Signal · Aug 18, 2026 · Technical Release · Source: OpenAI Blog · Impact: 4/5 · Sentiment: Positive
Large Language Models (LLM) & AI Market: Asana removed Enzyme in two weeks using Codex
Asana used OpenAI Codex to remove its outdated Enzyme testing system in roughly two weeks—work it had previously estimated would take five years. Engineers ran up to four parallel coding agents from a short prompt, with humans reviewing and approving changes twice daily. Model and infrastructure costs were about $12,000 versus an estimated $6 million and multi-year staffing plan. The project freed Asana to consider other long-running migrations, rewrites, and performance improvements that were previously considered impractical.
Major platform (OpenAI) case study showing LLMs and agent workflows can dramatically reduce engineering time and cost, signalling broader operational automation potential for software teams.
Key Takeaways & Evidence Grounding
- Asana removed the Enzyme testing tool after approximately 1.5 weeks of engineering effort spread over two calendar weeks.
- Asana used OpenAI Codex (powered by frontier models) with up to four parallel coding agents to perform the migration.
- Model and infrastructure costs for the project totaled about $12,000.
- Asana's previous plan estimated the same work would take about five years and cost roughly $6 million.
- Engineers reviewed every proposed change and checked progress twice a day; simpler prompts worked better.
Connected Companies & Entities
2 Entities mappedAsana
Work management software for teams, enterprises, and government agencies.
“In about two weeks, Asana completed work it expected to take five years....”
OpenAI
Foundation model company selling AI software, APIs and subscriptions.
“people use OpenAI Codex, powered by frontier models, to tackle large codebase changes...”
Ontology Mapping & Concepts
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.