OpenAI published a customer story on November 12, 2026, in which Amplitude estimated that its Wave product agent can help teams ship product improvements in 1-2 days, down from roughly a month (OpenAI). Amplitude describes Wave as roughly 80% powered by OpenAI models, including GPT-6.1 Sol, GPT-6 Luna and GPT-Image-2.5 Flare (OpenAI). The company calls Wave a proactive product agent that identifies what to build, hands work to coding agents and measures the impact (OpenAI).

Key Facts

ItemDetailSource
ProductWave, Amplitude’s proactive product agentOpenAI
Model shareRoughly 80% powered by OpenAI models, per AmplitudeOpenAI
Time to shipEstimated 1-2 days, down from roughly a monthOpenAI
Homepage test160% lift in card-driven sends versus control after one dayOpenAI
AvailabilityCustomers use Wave in betaOpenAI

Timeline

  • 2026-11-12: OpenAI published the Amplitude customer story describing Wave, its model mix and its reported results (OpenAI).

Details (as Reported)

Wave then brings together Amplitude data, session replays, customer feedback and code to identify and prioritize opportunities (OpenAI). When an opportunity is ready, teams can send it to coding agents, including Codex, with a technical plan and acceptance criteria (OpenAI). Amplitude’s own page describes the loop as designing new builds, orchestrating coding agents to generate pull requests and running experiments to validate impact (Amplitude (official statement)).

In a sample of 100 session replays, 80 users typed their own questions and only 12 selected a suggestion card (OpenAI). Wave proposed personalizing the cards, and after a brief review the team handed the work to a coding agent and A/B tested the change (OpenAI). In the readout Amplitude shared, messages sent through the cards rose 160% relative to control after one day (OpenAI).

For the Opportunity Investigator, GPT-6.1 Sol scored 4.06 out of 5 overall, against 3.62, 3.52 and 2.94 for three alternative models, and had the fastest median investigation time at 54 minutes (OpenAI). In the Data Science Agent evaluation, GPT-6.1 Sol scored 67.5 out of 100 against 69.3 for another leading model, at a mean model cost of $1.28 versus $2.63 per dispatch (OpenAI). GPT-6 Luna handles frequent, lower-cost tasks such as scanning code and running small evaluation checks (OpenAI). For solution wireframes, OpenAI’s image model scored 80.3 out of 100 against 63.9 for another image model across 15 opportunities (OpenAI).

Amplitude said most teams send bug fixes straight to coding agents, while new features and bigger bets get a human review (OpenAI). Among customer-launched experiments, flags and guides, one in four went live the same day Wave surfaced the opportunity, half in under four days and two in three within a week (OpenAI). Amplitude’s page cites a PixAI case in which an out-of-credits tooltip led to publishing rising nearly 10% and Day 1 retention rising 3% over control (Amplitude (official statement)).

Amplitude lists requirements for Wave: an eligible Growth or Enterprise plan, meaningful Product Analytics and Session Replay data, and a business goal to improve (Amplitude (official statement)). It also states that connected code is not used to train shared models (Amplitude (official statement)).

What Is Not Yet Confirmed

  • Amplitude states that demonstrating business impact depends on how quickly a team ships a change and collects enough data to measure its effect (Amplitude (official statement)).