How Amplitude ships product improvements in 1–2 days with OpenAI

Amplitude’s proactive product agent, Wave, uses OpenAI models including GPT‑6.1 Sol, GPT‑6 Luna, and GPT‑Image‑2.5 Flare to transform product insights into actionable plans. Wave integrates Amplitude data, session replays, customer feedback, and code to identify opportunities, prioritize fixes, and coordinate implementation with teams or coding agents like Codex. After identifying an opportunity, Wave generates a technical plan with acceptance criteria and can route work directly to coding agents for execution. The agent then measures outcomes and applies learnings to refine future improvements.

Wave reduced the time to ship product improvements from roughly a month to 1–2 days in Amplitude’s workflows. On Amplitude’s homepage, Wave identified that users bypassed generic suggestion cards and proposed personalizing them using recent user context. After implementation via a coding agent and A/B testing, messages sent through suggestion cards increased 160% relative to control within one day.

Amplitude evaluates models for Wave based on quality, latency, and cost, assigning tasks to specialists like GPT‑6.1 Sol for deep investigations and GPT‑6 Luna for frequent, low-cost checks. GPT‑6.1 Sol scored 4.06 out of 5 in Opportunity Investigator evaluations, outperforming alternatives on corroboration, false positive avoidance, and evidence alignment, with a median investigation time of 54 minutes.

Customers in beta use Wave to analyze session and analytics data, receive recommendations with technical plans, and control automation levels. Among customer-launched experiments, one in four went live the same day Wave surfaced the opportunity, half within four days, and two in three within a week. Amplitude aims to enable self-improving products by delegating more investigation and implementation to agents as OpenAI models advance.