Show Notes
When Tim Freestone first logged into ChatGPT on November 23, 2022, he turned to his wife and said, "Okay, this is a thing." Two and a half years later, he's the Chief Strategy Officer at Kiteworks, a PE-backed unicorn protecting how data moves in and out of the world's most regulated companies. This episode is part jazz appreciation, part AI philosophy, and part hard-earned playbook for any founder staring down the agentic era wondering whether their data exposure is about to catch up with them.
Tim's path is the kind founders should pay attention to. He spent the early part of his career writing grants for a performing arts college, then bootstrapped a New York marketing agency from zero to fifty employees and nearly ten million in revenue across a decade. The throughline was always building systems, and when AI collapsed the gap between intent and outcome, he went all in. A year ago he didn't know what a CLI was. Now he has more terminal tabs open than browser tabs.
Kiteworks itself is a study in repositioning. The company spent fifteen years as Accellion, a secure file transfer business that had commoditized into a struggling thirty-million-dollar revenue line. Then current CEO Jonathan Yaron, a veteran of Israel's elite 8200 unit, saw signal where others saw stagnation. He expanded the platform to cover every channel through which data enters and exits an organization: file share, email, managed file transfer, APIs, secure protocols. Tim arrived as CMO five years ago, recognized the brand confusion between Accellion and its Kiteworks platform, and convinced Yaron to elevate the product name to the company name. The rebrand stuck. The vision expanded. And now, in the age of agents, that same control plane is being extended to govern how AI systems access and move enterprise data.
The Intent-Data Layer Framework
SaaS historically sat as a complex translation layer between human intent and dataEntire job titles formed around mastering specific software stacks (Salesforce admins, etc.)AI strips out the complexity layer entirely, allowing natural language to bridge intent and data directlyThis democratizes data leverage for both good actors and bad actorsThe strategic implication: protection must move down to the data layer itself, not the software layer
The Control Plane for Data Model
Traditional security stacks at the perimeter, cloud, and endpointAll of those layers exist to protect data, but none control data directlyKiteworks operates at the data layer, mapping individual assets to individual agentsYes/no permissions on access, sharing, and use, asset by asset, agent by agentThis becomes the matrix companies need to maintain compliance in agentic workflows
The Regulator Doesn't Care Principle
Data exposure penalties apply regardless of cause: human error, agent action, orangutan typingPII, PHI, and CUI regulations remain in force even as agent regulations lagCompanies will face audits in 12+ months on agent activity happening todayInsurance policy: instrument controls now, before the legislative wave catches up
The Failure-as-Muscle Framework
Failures should be encouraged the way muscles must be pushed toward failure to growInsecure leaders pour gasoline on others' mistakes to distract from their own gapsStrong organizations normalize mistakes as part of the operating systemMentorship is less about seeking mentees and more about transparently sharing the lessons that informed every current decision
https://www.linkedin.com/in/freestone
https://www.linkedin.com/in/estesryan/