Marketo co-founder Jon Miller rethinks B2B marketing automation for the AI age

Jon Miller explains how Phave uses playlists to power one-to-one marketing.

Jon Miller, co-founder of Marketo, is returning to the category he helped create to build a marketing automation platform (MAP) for the AI age. 

Today, Miller and his team officially launched Phave, a new AI-powered platform designed specifically for enterprise B2B marketing organizations. After operating in stealth mode for two years, Miller told MarTech the platform now powers marketing operations across 10 companies. 

For Miller, Phave represents a deliberate attempt to rethink a category he thinks is falling behind the realities of modern revenue teams. Marketing automation platforms originally promised to help B2B companies identify, nurture, and convert prospective buyers at scale. 

Over time, however, the model degraded into rigid, rules-based systems that marketing teams too often use to subject prospective buyers to endless email streams while failing to account for the actual complexity of B2B purchasing decisions. 

Miller said the reality of modern B2B buying is clear: buyer behavior evolved, but legacy marketing automation stood still. 

“I like to say rules are good at what must be true, but they can’t handle ambiguity, and they can’t provide judgment about what is best,” Miller told MarTech. “And the reality is B2B buying is ambiguous and complex, and doesn’t lend itself very well to rules.”

B2B buying moved on 

The concept for Phave grew out of Miller’s experience at Engagio, the account-based marketing platform he founded after Marketo and sold to Demandbase in 2020. 

B2B marketers repeatedly shared the same frustration: legacy platforms could not keep up with operational shifts. Buying groups now dictate decisions, buyers conduct research anonymously, and marketing teams now manage the entire lifecycle through post-sale retention. 

Despite these changes, core MAP architecture remained anchored to static leads, linear campaigns, and manual rules. In Miller’s view, marketers tolerated these outdated setups not out of satisfaction but because no viable alternative existed. 

That operational friction shaped Phave’s architecture. Rather than building another point solution to stack on top of a legacy platform, Miller’s team spent its stealth period engineering a full enterprise replacement. 

Phave treats individual contacts, target accounts, and multi-person buying groups as first-class objects, each with its own intent scores and tailored buyer journeys. 

10X your SEO with Semrush for Enterprise.

The world’s most powerful SEO platform, purpose-built for Enterprise.

Request demo

Replacing rules with reasoning 

Phave’s biggest departure from legacy MAPs lies in its decision-making engine. Traditional marketing automation relies strictly on static logic: If a prospect does X, trigger Y. If their score hits a threshold, route them to sales.

But Phave replaces rigid triggers with AI reasoning models to handle scenarios where answers are far less predictable, which Miller says makes Phave better equipped to handle the inherent ambiguity in B2B buying. 

Rather than forcing marketing teams to manually configure every edge case into a rules engine, Phave reasons through available context to determine the logical next step. 

From campaigns to ‘playlists’ 

A view of the Phave Playlist feature. Click to enlarge.

Phave also introduced a structural shift in execution through what Miller, borrowing from the personalization of music platforms, calls a “playlist.” 

“The campaigns are the albums and the songs are the tactics, and we’re not using AI to necessarily spin up a new tactic for each person, but what we’re doing is we’re mixing them in the optimal order for each person,” Miller said.

Phave’s playlists continuously evaluate context to decide which “song” to play next based on real-time data across the individual, account, and buying group. 

Phave’s execution adapts dynamically as the system learns. When a buyer consistently opens emails on Tuesdays, for example, but skips webinars, Phave registers those preferences and automatically adjusts future outreach. 

Miller said this structure also resolves a chronic headache for marketing operations: competing campaigns. Conventional platforms allow separate teams to inadvertently email the same audience. Phave acts as a centralized traffic controller. If, for example, a high-priority product launch needs to reach an account on Wednesday, the platform automatically reschedules an overlapping webinar invite to Friday.

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

Miller said marketing operations leaders can build custom “skills” that instruct AI agents on organizational standards. These instructions capture precise campaign naming conventions, required UTM structures, and operational workflows. I

Instead of requiring team members to manually apply these guidelines every time, the agent executes the task according to established best practices. Miller said early Phave adopters report building campaigns two to three times faster.