Asana is showing how quickly AI agents are becoming more practical for everyday business work. In browser-agent tests using OpenAI’s GPT-6 Astra in Codex, the company reported a striking 76x reduction in model costs alongside a 5x speedup.
That combination matters because browser agents can help automate complex, multi-step tasks across web apps. By making these agents faster and far cheaper to run, Asana could offer customers more capable AI features while keeping performance smooth and scalable.
Why this is a win
- Major efficiency gain: 76x lower costs can change the economics of deploying AI agents.
- Better user experience: A 5x speed improvement means agents can feel more responsive and useful.
- Enterprise-ready progress: The results point toward AI tools that are both powerful and affordable enough for real workplace adoption.
While these results come from tests, they are a promising sign for the next phase of AI in productivity software: agents that can take action in the browser, complete tasks faster, and deliver value without runaway compute costs.