Amplitude said its Wave product agent can turn product insights into shipped changes in 1-2 days, down from roughly a month, by handing technical plans and acceptance criteria to coding agents including Codex.
On Amplitude’s homepage, Wave reviewed 100 session replays, found 80 users ignored generic suggestion cards, and proposed personalized cards based on recent charts, dashboards or referral pages.
That change lifted messages sent through homepage suggestion cards by 160% versus control after one day in an A/B test, while preserving free typing and fallback generic suggestions.
Amplitude said Wave is about 80% powered by OpenAI models: GPT-6.1 Sol for deeper investigations, GPT-6 Luna for cheaper high-volume checks, and GPT-Image-2.5 Flare for mockups.
Among customer-launched experiments, flags and guides, Amplitude said one in four went live the same day Wave surfaced the opportunity, half within four days and two-thirds within a week.
Could automating product changes in a single day lead to massive feature bloat rather than genuinely improving the user experience?
If Wave feeds user session data directly into OpenAI models, how can companies guarantee sensitive customer information remains completely secure?
If AI agents can identify problems, write technical plans, and deploy code autonomously, what exactly remains of the traditional product manager's role?