Amplitude says OpenAI-powered Wave ships product changes in 1, 2 days
Amplitude says its Wave product agent can help teams move from product insight to shipped improvement in 1, 2 days, down from roughly a month. The agent uses OpenAI models, including GPT‑6.1 Sol, GPT‑6 Luna and GPT‑Image‑2.5 Flare, to investigate opportunities, plan implementations and measure results.
- Amplitude says Wave can help ship product improvements in 1, 2 days.
- Wave combines product data, session replays, feedback and code.
- The agent can send technical plans to coding agents including Codex.
- A homepage test increased suggestion-card messages by 160% relative to control.
- Teams can review or automate changes based on their preferred workflow.
From product data to implementation plans
Wave brings together Amplitude data, session replays, customer feedback and code. Teams provide a product area and the goals or metrics they want to improve, and the agent identifies and prioritizes opportunities before developing proposed solutions.
When an opportunity is ready to build, Wave creates a technical plan and acceptance criteria that teams can send to coding agents, including Codex. After implementation, it tracks the outcome and uses the results to inform the next round of improvements. According to Amplitude, Wave is roughly 80% powered by OpenAI models.
Amplitude uses different models for different parts of the workflow. GPT‑6.1 Sol supports the opportunity manager and deeper investigations into code and data. GPT‑6 Luna handles frequent, lower-cost work such as scanning code and running small evaluation checks. GPT‑Image‑2.5 Flare generates product mockups from screenshots taken from session replays.
A homepage test increased card-driven messages
Amplitude tested Wave on its own homepage, where suggestion cards were designed to help users start a conversation with an AI agent. In a sample of 100 session replays, 80 users typed their own questions, while 12 selected a suggestion.
Wave recommended personalizing the cards with recent charts, dashboards or the page a user had visited. It also produced an illustration and technical plan that preserved free typing and displayed generic suggestions when personalization context was unavailable. After a brief review, Amplitude sent the work to a coding agent and ran an A/B test.
In the experiment readout shared by Amplitude, messages sent through the homepage suggestion cards rose 160% relative to the control after one day. “AI can now find useful work we never thought to assign it,” Amplitude CEO Spenser Skates said in the OpenAI case study. “OpenAI’s models can investigate a problem, write and run code, and use the results to decide what to try next.”
Amplitude matches models to the work
Amplitude says it evaluates quality, latency and cost for each task before assigning a model. For Wave’s Opportunity Investigator, GPT‑6.1 Sol scored 4.06 out of 5 overall, compared with 3.62, 3.52 and 2.94 for three alternative models. It led across all five investigation criteria and had a fastest median investigation time of 54 minutes.
In a separate evaluation of Wave’s Data Science Agent, GPT‑6.1 Sol scored 67.5 out of 100, compared with 69.3 for another leading model. Its mean model cost was $1.28 per dispatch, compared with $2.63 for the other model. Amplitude also said GPT‑6.1 Sol scored higher on tool-contract compliance, estimation and uncertainty, and evidence grounding.
For wireframes and mockups, OpenAI’s image generation model scored 80.3 out of 100, compared with 63.9 for another image generation model. The evaluation covered 15 opportunities and assessed communication quality, product fidelity, instruction adherence, text and UI legibility, and visual coherence.
Teams decide what reaches production
Customers across industries are using Wave in beta, according to Amplitude. The company says teams can choose how much automation to allow. Most teams send bug fixes directly to coding agents, while new features and larger bets receive human review.
Among customer-launched experiments, flags and guides, Amplitude says one in four went live the same day Wave surfaced the opportunity, half went live in under four days, and two in three within a week. The source says these figures were interpolated from supplied milestones and are illustrative rather than individual observations.
Amplitude’s approach positions agents as a link between product analytics and software delivery, while leaving teams responsible for reviewing larger changes. Skates said the company’s ambition with OpenAI is “to make self-improving products the way companies build software.”

