Kana's $15M Agent Play: Can Rapt/Krux Veterans Actually Fix Marketing Automation?

Remember when every marketing tool promised to be the "one platform to rule them all"? Yeah, how'd that work out?

Tom Chavez and Vivek Vaidya certainly remember. After selling Krux to Salesforce for $700 million in 2016, they've watched the martech landscape fragment into an even messier constellation of tools. Now they're back with Kana, a $15 million bet that AI agents can finally cut through the chaos.

But here's what makes this different from the usual "AI will save marketing" pitch deck fever dream.

Beyond the Chatbot Theater

Most marketing AI today is glorified Q&A. Ask it something, get an answer, manually implement the suggestion. Rinse, repeat, wonder why you're paying enterprise fees for a fancy search bar.

Kana's agents supposedly execute multi-step workflows autonomously. Think: monitoring campaign performance, identifying underperforming segments, automatically adjusting targeting parameters, then generating explanatory reports. Without human babysitting at every step.

That confidence comes from experience. Their previous ventures include:

• Rapt: Consumer behavior analytics (acquired by Microsoft, 2008)
• Krux: Data management platform (the $700M Salesforce payday)
• super{set}: The startup studio that incubated Kana for 9 months

Two successful exits in martech isn't luck. It's pattern recognition.

The "Build With" Gambit

Here's where Kana gets interesting. Instead of the typical "build or buy" enterprise software dance, they're pushing a "build with" model. Human-in-the-loop oversight, marketer approval for agent actions, continuous feedback loops.

Smart positioning. Nobody wants autonomous AI agents burning through ad budgets without guardrails. But the promise of customizable workflows that actually adapt? That hits different.

The synthetic data generation component is equally clever. Why pay premium rates for third-party audience data when you can generate synthetic datasets for testing? Faster experimentation cycles, lower costs, fewer vendor dependencies.

Translation: They're attacking the martech stack from multiple angles.

The Timing Question

Here's my cynical take: Is this the right moment, or just convenient timing?

Enterprise buyers are definitely moving beyond basic chatbots. The "AI agent" category is having its moment. Investors are throwing money at anything that promises autonomous task execution.

But marketing automation has been promising flexibility for decades. Remember when Marketo was going to democratize sophisticated campaigns? How about when every CDP claimed it would unify your data stack?

Sure, Tom. The market is always crying out for something.

Hot Take: This Might Actually Work

Here's my controversial opinion: Kana has better odds than most.

Not because their AI is necessarily superior. Not because marketing automation suddenly got easier. But because they've built and sold martech platforms before. They understand enterprise sales cycles, implementation challenges, and the gap between demo magic and production reality.

Their 25+ years of combined experience matters more than whatever LLM they're fine-tuning. They know which promises to make and which corners to avoid cutting.

Plus, targeting smaller companies first is smart. Enterprise marketing teams have dedicated specialists and existing tool investments. Mid-market companies? They're drowning in point solutions and desperate for something that actually works together.

The $15 million seed round gives them runway to prove the concept without premature enterprise complexity. If the agents deliver even 70% of the promised flexibility, they'll have customers.

My prediction? Either they get acquired by a major martech player within two years, or they become one themselves. The founders' track record suggests they know which outcome pays better.