Amplitude unveils autonomous AI agents for analytics

Amplitude has launched a set of AI agents that continuously monitor product usage, flag significant changes in key metrics, and recommend actions in real time.

The release includes a Global Agent and four specialised agents for dashboard monitoring, session replay analysis, experimentation, and user feedback. Amplitude positions the tools as a response to faster software delivery cycles, where teams can ship features more quickly than they can assess their impact.

Amplitude is best known for behavioural analytics used by product, marketing, and data teams. More than 4,500 organisations use its platform, including Atlassian, Burger King, NBCUniversal, Square, and Under Armour.

The agents are designed to reduce manual effort in investigating metric changes, diagnosing performance issues, and identifying optimisation opportunities. Amplitude is also rolling out updates based on the Model Context Protocol (MCP) to support integrations with third-party tools used across product development and collaboration workflows.

The Global Agent spans Amplitude's analytics environment. It answers plain-language questions using behavioural data in the system, and can also build dashboards, investigate root causes across funnels and segments, and explain what is driving changes in product performance.

Specialised agents

The Dashboard Monitoring Agent detects meaningful metric changes over short time frames, investigates what has shifted, and can send insights and recommended actions through Slack or email.

The Session Replay Agent reviews user sessions continuously and highlights friction points. It can also quantify revenue impact and recommend fixes.

The Web Experimentation Agent supports testing workflows by designing and launching experiments, analysing results, and recommending whether to roll out or iterate. The process keeps a human in the loop.

The AI Feedback Agent processes unstructured input such as surveys and support tickets, and maps themes to observed user behaviour in Amplitude.

Amplitude says its approach differs from AI tools that rely solely on a data warehouse. Its agents operate inside its behavioural analytics product, which it says provides context on users, journeys, funnels, and cohorts rather than single-query answers.

Workflow integrations

The MCP updates aim to bring Amplitude data into the tools teams already use for analysis, development, and collaboration. Amplitude listed integrations with Claude and ChatGPT for summarising behaviour and answering product or campaign questions using Amplitude data.

For software development workflows, product context can be surfaced inside Cursor and Claude Code, and behavioural context can be added to GitHub pull requests. Amplitude also referenced AWS Kiro as a place where teams can build and ship AI features with integrated product insights.

In design and prototyping, Amplitude said integrations cover Lovable and Figma Make. For product and collaboration, it cited Notion agents for analysis and making Amplitude charts searchable in Atlassian Rovo. It also referenced Outreach's use of behavioural signals in sales and engagement activities.

Amplitude framed the launch against the rapid growth in AI coding assistants, naming companies including Anthropic, OpenAI, Cursor, and Lovable. It argues that analytics has become a bottleneck as release cycles compress.

"We're entering a new era of analytics-one where AI can monitor your product around the clock, and free up your team to focus on improving the experience," said Spenser Skates, Co-Founder and CEO, Amplitude. "Today, we're launching the first fully autonomous analytics agent. It's going to reinvent how product decisions get made."

Early users

Several customers and partners shared how they expect to use the tools. NTT DOCOMO said it uses Amplitude for self-serve analytics at scale across its organisation.

"Amplitude has helped NTT DOCOMO scale self-serve analytics to more than 1,000 active users and significantly reduce the time required to analyse campaign effectiveness," said Takashi Suzuki, SVP and GM of the Data Platform Department, NTT DOCOMO. "With Amplitude AI Agents, our teams can streamline analysis directly from existing dashboards, helping us move faster while improving conversion rates and lowering cost per acquisition."

Mercado Libre said expanding analytics usage across the organisation required more than access to data.

"Increasing our users required more than just access to data. It required structure and automation," said Matias Caratti, Product Shipping Supervisor, Mercado Libre. "Dashboards provided a single source of truth, and AI Agents enabled us to find data on our own. We didn't have to rely on manual reports, and we were provided with automatic insights on funnel performance, countries with the best conversion, and fluctuations in contact rate."

Cursor, which Amplitude also cites as an integration target, said the MCP work brings user insights into the development environment.

"Amplitude MCP and Skills bring user insights directly into agent context in Cursor," said Joshua Ma, Engineering Lead, Cursor. "This allows teams to quickly ship features, measure impact, and build smarter experiments for the next release."

Amplitude said agents can take action within its platform after analysis and expects teams to use them for continuous monitoring and recurring workflows, such as trend reviews, session analysis, and experiment evaluation.