Operate The Predactiv Data Platform from the AI tools you already use, like ChatGPT and Claude, or build its capabilities directly into your own product, going from a plain-language request to a live, activation-ready audience.

What we’re introducing

Two new ways to reach The Predactiv Data Platform, in addition to the Predactiv interface:

  • Operate from your AI tools (MCP Server): Connect The Predactiv Data Platform to generative AI tools like ChatGPT, Claude, Cursor, and Manus through the open Model Context Protocol (MCP), then build, plan, and activate audiences in natural language, no dashboard required.
  • Embed into your own product (Platform API): Build Predactiv’s capabilities directly into your own products and workflows through a standards-based API and token-based authentication, with no custom infrastructure to stand up.

Both run on the production Predactiv Data Platform and honor the same governance, with feature flags and organization-level entitlements respected on every request.

What you can do

  • Build and preview audiences just by describing them in plain language, with live size estimates before activation
  • Onboard and manage first-party datasets, ready for audience creation
  • Pull audience insights and analysis on demand
  • Deliver audiences to the destinations you already use, such as DSPs, SSPs, clean rooms, and identity partners
  • Discover and apply predictor signals and demographic and geographic filters

Connecting The Predactiv Data Platform to a generative AI tool does more than retrieve data. A user might pull the insights for two audience segments, ask which is sharper and why, turn the answer into an executive-ready summary or chart, then bring in public web research to benchmark the segments and surface new signals that expand the stronger one, all in the same conversation.

These capabilities pair The Predactiv Data Platform’s own audience creation, sizing, insights, and activation with the AI tool’s reasoning and reach across the open web, shortening the distance from audience discovery to activation and letting teams build their own audience intelligence layer on top of the platform’s data.


The same platform, two ways to work with it.

From an AI tool (MCP)

ChatGPT
Claude
Cursor
Manus

For teams that already work in AI tools. Connect Predactiv, then build, size, and activate audiences just by describing what you want, no dashboard required. For example:

Building and sizing a Predactiv audience from a prompt in Claude

Predactiv builds the audience, returns a live size estimate, and can deliver it to your destination, all in the same conversation.


From your own product (Platform API)

For engineering teams embedding Predactiv into their own product. Authenticate once with OAuth 2.0, then call the same audience and dataset capabilities directly from your code. For example:

The same audience and dataset capabilities, on the same platform and governance as the Predactiv interface. Illustrative, your team builds against the full API documentation.


Open by design

Built on open standards, so you decide how Predactiv fits into your stack, not the other way around.

Works with what you use

Activate to the DSPs, SSPs, clean rooms, identity partners, and destinations you already rely on, with no proprietary workflow to adopt.

Enterprise-grade governance

Privacy-centric architecture, with the same controls, feature flags, and entitlements as the rest of the platform.