Marketo Co-founder Launches Phave Platform for the AI Era

Despite these shifts, MAP architecture remained tied to static leads, linear campaigns and manual rules. In Miller's view, marketers tolerated outdated systems not out of satisfaction, but for lack of a viable alternative.

This operational pain defined Phave's architecture. Instead of another point solution layered on top of legacy platforms, Miller's team built a full enterprise replacement during the stealth development period. Phave works with individual contacts, target accounts and buying groups as first-class objects — each is assigned its own intent scores and individual buyer journeys.

Rules Are Replaced by Reasoning

Phave's key difference from classic MAPs is its decision-making engine. Traditional automation relies on static logic: "if a prospect did X, trigger Y; if their score crosses a threshold — hand off to sales." Phave replaces rigid triggers with AI reasoning models that work in scenarios without predictable answers. Instead of manually configuring every edge case in a rules engine, the platform analyzes the available context and determines the logical next step.

From Campaigns to "Playlists"

Miller calls the structural shift in execution a "playlist," drawing an analogy with music service personalization.

"Campaigns are albums, and songs are tactics. We don't use AI to necessarily create a new tactic for each person — we remix them in the optimal order for each," Miller explained.

Phave's playlists continuously assess context and decide which "song" to play next, based on real-time data about the buyer, the account and the buying group. If a buyer consistently opens emails on Tuesdays but skips webinars, the platform registers the preferences and automatically adjusts future touches.

This structure also solves a chronic marketing operations pain — competing campaigns. Conventional platforms allow different teams to unknowingly send emails to the same audience. Phave acts as a central dispatcher: if a high-priority product launch must reach an account on Wednesday, the platform itself moves an overlapping webinar invitation to Friday.

The platform was designed for an era when marketers interact with software primarily through AI agents rather than manual dashboard navigation. Phave exposes 319 tools via the Model Context Protocol (MCP), allowing external agents and internal tools to trigger actions directly.

According to Miller, marketing operations leaders can create custom "skills" that set organizational standards for AI agents: exact campaign naming conventions, mandatory UTM structures, operational processes. The agent executes tasks according to established best practices without manually applying guidelines. According to Miller, early Phave users assemble campaigns two to three times faster.