RudderAI brings agentic power to the entire customer data lifecycle

The future is in the hands of teams that build with AI, and RudderStack is evolving to equip the builders. We’re empowering more teams, both technical and non-technical, to build agentic workflows and applications with their customer data.

Today, we’re launching RudderAI, RudderStack’s agentic layer, anchored by our CLI and MCP, to bring agentic power to the entire customer data lifecycle.

RudderAI gives data and engineering teams a natural language control plane to operate RudderStack, and it brings business teams closer to customer data than ever. Now marketers can analyze data to understand users, uncover high performing segments, build audiences, and sync them downstream from one seamless, agentic workflow.

Next, we’re unlocking long running, autonomous agents to safely act on your customer data (and the future is already in preview. Keep reading to the end).

A generational shift for the customer data platform

Yesterday’s CDP was UI-locked, highly opinionated, and human configured. Platform value was in the opinion, the workflow, and the interface because humans did all the configuration by hand. But agents have flipped that value proposition on its head and ushered in what Martech analyst Scott Brinker calls the move from apps to infrastructure as teams build custom agents and applications that can outperform opinionated tools.

What matters now, Brinker argues, isn’t so much the interface as the foundation underneath: “the infrastructure that lets all those apps and agents work together safely, coherently, and at scale. Data. Identity. Orchestration. Governance. Observability. Semantics. Context.”

That foundation is what we’ve been building since day one. Our driving principles and core products have always been aligned with the needs of the AI era:

• Warehouse-native DNA: RudderStack is built with a modern architecture at its core. Instead of storing customer data in a black box, it supports deep, seamless warehouse integration to move data in and out. Plus, it can operate on data in the warehouse to stitch identities and model business logic, enabling teams to establish their warehouse as the central, open hub to fuel analytics, activation, and AI.
• Governance included: RudderStack brings event data governance controls together in a central command center with deep observability and fine grained control over data schemas, validation logic, consent management, and PII. This gives teams maximum control over data quality and compliance and supports seamless workflows.
• Code-based power: RudderStack is built for data teams with programmatic interfaces and APIs to enable code-based management of the customer data lifecycle and provide a foundation for agentic interaction.

RudderAI isn’t a pivot. It’s a natural evolution for RudderStack, and it represents the next generation of the CDP, one where humans direct agents, and agents do the work. This paradigm is already driving transformation for our customers.

"With RudderStack the data team is no longer a bottleneck," says QJ Flores, Staff Data Engineer at VSCO. "We built a tracking agent with Cursor that reduced time from data request to insight from six weeks to a few days." That's roughly 15x faster time to insight thanks to agentic workflows built on top of RudderStack agent ready infrastructure.

Agentic power for the entire customer data lifecycle

RudderAI leverages RudderStack’s IaC foundation along with our MCP to expose platform wide capabilities to agents, enabling AI native workflows across the customer data lifecycle. And you can take advantage of it with the tools you already use, Claude, Codex, Cursor, or any MCP-enabled agent.

For data and engineering teams, this means you can direct agents to do configuration work and debug issues with natural language. For business teams, it removes the data team as a bottleneck, empowering them to explore customer data to understand behavior and then activate it to make use of their insights. Human-directed, agent-executed actions from collection to activation. Here’s what you can do with RudderAI:

Collect: Automatically generate tracking code that fits your tracking plan and codebase conventions. Configure, manage, and debug your customer data pipelines, and write transformations to shape data in flight with agentic workflows.

Unify: Use guided, agentic workflows to resolve identities, create customer 360 profiles from all the data in your warehouse, build models for predictive features, and create a data graph to make warehouse data self-serve ready.

Activate: Marketers can explore and analyze complete customer data to uncover high performing segments, build audiences, and sync them to all their tools using natural language from a single agentic workflow. Product teams can analyze user journeys, build funnels, and share dashboards with natural language.

Govern: Have agents create a central data catalog to define every event you track then build tracking plan rules on top so only activation ready data makes it downstream.

To make building easier, we're introducing RudderStack Agent Skills, a Claude Code plugin that teaches Claude how to drive every programmatic RudderStack surface: the CLI, the MCP server, and Terraform. Each skill packages the domain knowledge and the workflow, the right preflight checks, commands, and recovery paths, for the work that agents could trip up on: designing event taxonomies, managing tracking plans and data catalogs, generating type-safe SDKs, and debugging instrumentation. Check it out on GitHub today.

One more thing: The future is here

What's next? Imagine if anyone on your team could query your warehouse, debug a broken funnel, and ship instrumentation, all in a single conversation. That's Lookout, RudderStack's new AI-powered analytics and instrumentation workspace. We're announcing it later this week. Stay tuned.

And we’re not stopping there. To enable every company to safely deploy autonomous agents on their customer data to accomplish any use case, just by describing their goals, we’re building RudderCrew. RudderCrew is our agent harness, a governed RudderStack runtime for autonomous agents that act on customer data.

Real-time decisioning is the first use case we’re tackling. Because, as Snowflake’s Florian Delval recently posited, decisioning is a customer data unsolved challenge with high upside that may be the next frontier for the customer data platform, and it’s a problem perfectly suited for agents. There are some existing solutions on the market, but none deliver the holy grail, and the promise CDPs have held out for ages: True 1:1 personalization. Real-time decisioning is already here in research preview. Contact us if you want to learn more.