Amplitude has launched Wave, a proactive product agent that spots what’s worth building, helps teams ship the change with coding agents, and then measures whether it worked. According to OpenAI, which published the case study, Wave is built mainly on OpenAI models and has helped Amplitude ship product improvements in one to two days.
That timeline is the headline. Going from “we should fix this” to “it’s live and we know if it worked” usually takes weeks, sometimes a full quarter. Amplitude is betting that one agent can cover that whole loop.
What Wave Actually Does
OpenAI describes Wave as covering three stages of product work. Here’s what each one means in practice.
- It finds what to build. The agent is “proactive,” so it doesn’t wait for someone to ask a question. Amplitude’s whole business is product analytics: tracking how users click, drop off, convert and churn. Wave sits on top of that behavioral data and flags opportunities on its own. Think of an analyst who never stops looking at your funnels.
- It helps ship the change. This step marks the real shift. Wave doesn’t stop at the insight. It works with coding agents to turn a finding into an actual product change. Analytics tools have always been good at telling you something’s broken. They’ve never helped you fix it.
- It measures the impact. After the change goes live, Wave tracks whether it moved the numbers. That closes the loop and feeds the next round of ideas. Here Amplitude’s existing experimentation and analytics stack gives it a clear edge.
Why This Matters
What stands out here is the move from dashboards to action. For years, analytics platforms delivered charts, and humans did the rest. Someone had to read the chart, write a ticket, wait for an engineering sprint, ship, and then remember to check the results. Every handoff added days.
Wave squeezes those handoffs into one agent-driven workflow. If the one-to-two-day cycle holds up beyond Amplitude’s own team, product teams could run far more experiments in the same amount of time. More experiments usually means more learning, and that compounds.
It also shows where enterprise AI is heading. The first wave of AI features in SaaS tools were chat boxes that answered questions. The current wave is agents that take actions across systems. Amplitude joining that shift makes sense, because it already owns the data that tells you what to fix.
How It Fits the Competitive Picture
The source doesn’t compare Wave directly with rival products. Still, the context is clear. Analytics competitors like Mixpanel, PostHog and Pendo are all adding AI features, and coding agents from OpenAI, Anthropic, GitHub and others are maturing fast. Amplitude’s pitch is that it connects the two: insight on one end, shipped code on the other, and measurement wrapped around both.
It also says something about OpenAI’s strategy. Showcasing partners like Amplitude helps OpenAI position its models as the engine for agentic workflows inside established software companies, not just inside ChatGPT.
Caveats Worth Noting
A few things to keep in mind before getting too excited:
- It’s a case study, not a benchmark. The one-to-two-day figure comes from Amplitude’s own experience as reported by OpenAI. Your results will depend on your codebase, your team and your review process.
- Availability and pricing weren’t detailed. The source doesn’t say who can access Wave, what it costs, or whether it’s tied to specific Amplitude plans.
- Humans still matter. Agents that suggest and ship changes need guardrails. Teams will want code review, approval steps and clear rollback plans before letting any agent touch production.
- Data quality is everything. An agent that acts on bad tracking data will confidently ship the wrong fix. Teams with messy event tracking should clean house first.
What Comes Next
Wave points to a future where the product loop (observe, build, measure) runs on autopilot for small improvements, while people focus on the bigger bets. The open question is trust. How much will product leaders let an agent decide and ship on its own?
Expect other analytics and product platforms to answer with their own end-to-end agents soon. The full case study is available from OpenAI for anyone who wants more detail on how Amplitude built Wave.