ARTICLE
As part of a recent Executive Exchange for C-level and IT leaders, Nutanix’s AI Leader Jason Langone spoke with Guy Kawasaki, Chief Evangelist at Canva and one of the technology industry's most recognizable voices on innovation and evangelism.
Kawasaki, who helped shape brand evangelism during the early days of the Macintosh and later returned to Apple as chief evangelist, discussed where agentic AI is heading and what it means for leaders trying to move fast without losing control. Several themes offer a useful lens for executives thinking about how to govern and scale autonomous AI enterprise-wide.
Kawasaki's central metaphor for disruption was ice. Think of Ice 1.0 as having to manually cut blocks from a frozen lake. Ice 2.0 was the ice factory and the ability to freeze water anywhere, anytime. Then came Ice 3.0 with the refrigerator, a personal ice factory in every home. Each phase wasn't simply a faster or cheaper version of the last; it was a fundamentally different curve altogether.
The same logic applies to agentic AI. Organizations that treat autonomous agents as a marginally better version of existing automation are optimizing the wrong curve. The bigger opportunity is recognizing when a technology enables an entirely new way of operating, and asking what becomes possible rather than what is just cheaper or faster.
Some of the most compelling moments in the conversation were the small, personal examples of agentic AI already at work, such as an agent that scans a calendar for flights and sends automatic status updates, one that reviews video and delivers a coaching report in under a minute, and another that manages routine family logistics without being asked twice. While these use cases might seem trivial individually, they illustrate where the technology is headed, and that’s agents that take initiative, monitor conditions, and complete multi-step tasks autonomously rather than simply answering questions.
The enterprise equivalent is what leaders now have to plan for — not a single chatbot bolted onto a workflow, but a mesh of autonomous agents operating across systems, requiring the same qualities that made those personal examples work: clear objectives, the right data access, and a way to verify the outcome.
Kawasaki drew a direct line back to Macintosh's early software strategy. Apple could have built every application itself, but instead it opened the platform and let outside developers innovate, a decision that produced categories nobody inside Apple had imagined, like desktop publishing. His argument for agentic AI is much the same. No single provider, however capable, can out-innovate an entire developer and vendor ecosystem.
Locking your organization into one model, platform, or provider creates two problems. First, switching to something better later becomes harder and more expensive, and second, you shut yourself out from adopting the next breakthrough if it comes from a competitor. For CIOs, that argues for infrastructure and architecture choices that preserve flexibility, so you can adopt the best available model or agent framework as the landscape shifts, rather than staying anchored to yesterday's decision.
Asked how leaders should balance boldness with responsibility, Kawasaki recommended a portfolio approach, where some initiatives are unproven and experimental, and others are reliable revenue generators funding that experimentation. A portfolio that is entirely speculative is reckless, but a portfolio that is entirely proven risks obsolescence without anyone noticing until it's too late.
Applied to agentic AI, this means treating autonomous agent deployment as a spectrum rather than an all-or-nothing bet. Mix mature, well-governed use cases in production with a smaller set of higher-risk pilots designed to find the next curve jump before a competitor does.
Perhaps the most direct takeaway from the discussion was Kawasaki's acknowledgment that caution, while understandable given shareholder pressures, can itself be the biggest risk. The companies most likely to miss the next curve are often the ones most successful on the current one, precisely because they have the most to protect and the shortest planning horizon.
The practical middle ground isn't reckless speed, it's building governance structures that make measured risk-taking possible, such as auditability, oversight, and guardrails that let you move quickly without moving blindly.
Kawasaki closed with a grounded reminder amid AI headlines predicting existential risk. For the overwhelming majority of people and organizations, the value of AI is far more mundane and far more real. It helps people write, research, plan, and produce, every single day. That practicality is worth holding onto. Governance and strategic ambition matter, but so does simply using the technology, letting it act as a research partner, a devil's advocate, and a productivity multiplier across your organization.
Together, these takeaways point to a common thread for any executive navigating the agentic shift. It’s critical to pair genuine ambition with real openness, fund experimentation without betting the business on it, and build enough governance to move fast responsibly.
The conversation with Guy Kawasaki was part of a broader executive exchange hosted by Nutanix exploring the evolution toward agentic AI and the leadership mindset required to harness it responsibly. As agentic AI matures, the organizations that get governance and scale right together will be the ones positioned to jump the curve rather than simply ride it.
Check out more articles on Nutanix Executive Focus.