Agents Overview

Amplitude AI Agents are specialized, goal‑oriented AI experts embedded in Amplitude’s platform. Unlike generic copilots, agents can access Amplitude analytics, session replay, experiments, surveys, and other data types to surface insights, investigate issues, generate hypotheses, and recommend actions.

How Agents work with Amplitude

Amplitude AI Agents bring together analytics, session replays, experimentation, guides, and survey data from across your account to build a deep understanding of user behavior. The more complete your Amplitude setup is, the more context and action recommendations the Agents have to work with.

The first Amplitude Agents are task-specific agents tuned to solve common problems like monitoring dashboards, bulk analyzing session replays, and optimizing conversion on websites.

Select a dashboard or define a goal to get started. Example goals include improving trial-to-paid conversion or increasing feature adoption. After you set the context and instructions, the Agent analyzes your data for patterns, identifies issues, and recommends next-step opportunities. It can also carry out actions you approve, such as running an experiment or sending targeted guidance to specific user segments.

Create and schedule multiple agents to identify key takeaways and improvement areas across your product. Agents push insights and actions into the places where you work, such as Slack, Amplitude, and email.

Use natural language to guide Agents, set custom instructions and context to improve accuracy and relevance, then review insights and approve recommendations. Amplitude manages the orchestration by combining models from OpenAI, Claude through AWS Bedrock, and Gemini with its own memory and tooling.

Example use cases

• Automate reporting with the Dashboard Agent.
• Discover behavioral insights with the Session Replay Agent.
• Explore strategies with the Website Conversion Agent.