AI-Native Marketing Automation Has Arrived: Phave Replaces Rules With Reasoning

On September 23, 2026, Jon Miller — the co-founder of Marketo and Engagio — brought Phave out of two years in stealth: an AI-native marketing automation platform positioned to replace Marketo, Pardot, Eloqua, and HubSpot. The pitch is blunt: AI decides what each person, account, and buying group should receive next.

This matters to anyone running business automation, not just marketers. The marketing automation stack was one of the earliest places where companies codified "if this, then that" workflows at scale. When the person who helped build that industry says rules were the wrong foundation, it is worth seeing what he built instead. (Marketo Co-Founder Jon Miller Announced the Launch of Phave - MarTech Cube)

What Phave actually launched

Phave's core is Maestro, an intelligence layer that computes a "Playlist" for each person: the touches most likely to reach the marketer's objective, ordered and timed for that individual. Before anything goes out, Maestro weighs everything else already queued for that person across every campaign. Campaigns are the albums — made and approved by the marketing team — and the Playlist mixes them into the right order for each listener.

Three design choices matter for anyone evaluating automation tooling:

1. Plain-language operation. Marketers describe what they want in plain language instead of translating it into configuration, so campaigns, audiences, scoring, and automations no longer get stuck behind whoever knows the platform.

2. Headless by default. Phave runs through its own interface, through the AI client a team already uses, or headlessly via MCP and a versioned REST API — with permissions and approvals enforced the same way whichever door a request comes through.

3. Accounts as first-class records. Accounts and buying groups each carry their own score and lifecycle stage. One account can hold several buying groups at once, at different stages — prospect and customer alike — which makes marketing to the committee that expands an account possible, not just the one that bought it.

The platform reached general availability in August 2026, with enterprise users including SambaNova, SPS Commerce, mabl, Servion, and Hypha. Bruce Eidsvik, Chief Growth Officer at Servion Global Solutions, said his team has more than doubled its marketing execution volume on the platform. (Yesterday's MarTech, AI & CX News | September 24, 2026 - The Agile Brand Guide)

Reasoning instead of rules: the core idea

Miller's thesis is clean enough to quote: "Rules are good at what is allowed, but bad at deciding what is best. Consent, frequency and quiet hours should be rules. Choosing which of two good emails a person should get was never a job for a rule. Reasoning models can make that choice now. The legacy platforms stopped innovating years ago, and marketers are ready for something built for how B2B works today."

The technical argument underneath is straightforward. Every decision a legacy platform makes is a rule somebody wrote in advance: who belongs in a segment, what a lead is worth, which email comes next. Rules cannot handle ambiguity or judgment, so every case has to be spelled out ahead of time — and most mature programs run two or three nurture tracks, not because their buyers are simple, but because that is all a rules system can manage.

This is the same design line the industry is converging on from several directions at once. It is also a claim worth testing rather than applauding — which brings us to the launch details that deserve scrutiny.

The launch details worth scrutinizing

The self-graded scorecard. Phave scored itself and the legacy platforms against 481 requirements drawn from twelve enterprise evaluations, with each requirement scored and weighted by Claude rather than by Phave. It reported 2.97 for itself against 2.46 for HubSpot and 2.38 for Marketo, matching or beating every evaluated platform on at least 90% of requirements, with the methodology published at phave.com.

Read the fine print the way a practitioner should: Phave picked the 481 questions, and Claude marked the answers. A published methodology is genuinely useful — it lets a buyer rerun the evaluation with their own requirement list and weights. But no one should budget against a vendor's self-assessment, and the same discipline applies to every platform already in your stack.

The pricing model. Phave starts at $36,000 a year, priced on the number of people who receive something in a given month rather than on database size. That is a notable shift for a category that has spent a decade charging for how much data sits in the system. Pricing on reach instead of storage aligns the vendor's incentives with actual execution — a trend worth watching across all your automation vendors.

The customer evidence so far. The launch's named customer result is a volume figure — doubled execution output — which measures throughput, not revenue impact. Throughput is a fine early signal for a platform built to remove configuration bottlenecks, but it is not proof of pipeline. Treat it as a reason to trial, not to migrate. (Yesterday's MarTech, AI & CX News | September 24, 2026 - The Agile Brand Guide)

What this means for your business

Whether or not you ever evaluate Phave, this launch puts the design split in writing, and it applies to every automated workflow your company runs:Draw the line between constraints and judgments. Write down which decisions are hard limits no model may cross — consent, frequency caps, quiet hours, pricing floors, brand and legal red lines — and which are judgments a model may make, like which message, which sequence, which timing. Rules are for what is allowed; reasoning is for what is best. If your current automation cannot express that split, it is a legacy architecture.

Audit your rule-based workflows for ambiguity. Anywhere your team maintains long exclusion lists or branching decision trees — workflows configured once and never touched again — you are paying the tax rules charge on ambiguity. Those are the first candidates for a reasoning-based approach, and also where silent failures hide.

Rerun vendor scorecards on your own requirements. Phave published its evaluation methodology — use it. Score any vendor, incumbent or challenger, against your own requirement list and weights before moving budget. A vendor's self-assessment tells you what they think matters; only your weights tell you what does.

Watch the pricing trend. Reach-based pricing instead of database-size pricing honestly aligns incentives. The next time any of your automation vendors renews, ask why you are paying for storage instead of outcomes.

Marketing automation was the proving ground for rule-based business workflows, and the person who built the old proving ground just declared it obsolete. The reasonable move is not to migrate tomorrow — it is to find which of your own workflows still run on rules written in advance for cases that never stay that simple.

Not sure where reasoning-based automation could replace brittle rules in your own stack? That is exactly what our free AI audit covers. Get your free AI audit — we will map your current workflows, flag where rules are failing you, and show you where AI reasoning can take over safely.