Asana's Browser Agent Saves Money by Pretending to Work

The claim: Asana announced it used GPT-6 Astra in Codex to make its Asana browser agent 76x cheaper and 5x faster in internal tests, aiming to provide customers with more capable automation models.

The reality: Saving money on a tool that still requires a human to log in, fix authentication errors, and manually verify every data entry in Chrome does not change a single enterprise workflow.

Why should I care? You shouldn't
Cutting the cost of a solution to a problem nobody asked you to automate is just efficient waste.

Read the original: Asana cuts model costs 76x in browser tests with GPT-6.1 Sol