Martech Veterans Raise $15M for AI Agent Startup Kana
Kana, a startup building AI marketing agents, emerged from stealth with $15 million in new funding from Mayfield. The company was founded by serial entrepreneurs who previously founded Rapt (acquired by Microsoft) and Krux (acquired by Salesforce). The raise demonstrates that strong founder pedigree and a focus on martech automation continue to attract significant early-stage capital.
- The founders of Kana are Tom Chavez and Vivek Vaidya; this is their fourth company together. Their previous startup, the data management platform Krux, was acquired by Salesforce for approximately $700 million in 2016 to enhance the Salesforce Marketing Cloud. - Their first company, Rapt, provided advertising yield management solutions for digital publishers and was acquired by Microsoft in 2008. The technology was integrated into Microsoft's Atlas Publisher Suite to compete with Google's DoubleClick and was already used by MSN, where it had reportedly increased ad revenue by up to 20 percent. - Kana was developed inside super{set}, an AI-native startup studio also co-founded by Chavez and Vaidya. The studio recently had another successful exit with the data collaboration platform Habu, which was acquired by LiveRamp in 2024. - The company's AI agents are designed to automate specific marketing tasks such as building media plans, generating reports, and optimizing campaigns by leveraging a company's own first-party data. This approach aims to replace rigid, traditional marketing automation tools with more flexible, customizable agents. - As part of the investment, Navin Chaddha, Managing Partner at Mayfield, will join Kana's board of directors. Mayfield has a history of investing in the marketing automation space, having previously backed Marketo, which later had a successful IPO. - Kana's platform focuses on "augmented intelligence," which allows users to approve, correct, and coach the AI agents' outputs before they are deployed. This keeps a human in the loop to ensure quality while still accelerating workflows.
