As Chief Product Officer at Treasure Data, Rafa Flores has spent 20 years across big tech as a self-described “fixer and builder.” It’s his unconventional journey with AI that’s rewriting how modern product teams operate.

After a freak accident left him in a coma for ten days and battling epilepsy, he turned to early AI and machine learning when traditional medicine fell short. By combining voice recognition, sleep data, and behavioral signals, he built a predictive model that could forecast seizures before they happened. This model gave him back his independence and two seizure-free years. This is the kind of user-driven innovation that every product leader talks about, but few experience. “I used my chops to develop something that could protect my life,” Flores explains.

AI-native products are built with AI from day one, while AI-enabled products retrofit AI into existing systems. The distinction matters because AI-native companies are scaling from zero to $100 million in recurring revenue in three months. At Treasure Data, Flores is navigating both worlds simultaneously, shipping three new product lines in three months while maintaining their foundational data platform.

The secret? Ditching rigid roadmaps entirely. “Your cycle is the hour that’s in front of you,” he says, acknowledging that waterfall project management and even agile methodologies are being shelved for real-time iteration.

The Treasure Data team does GTM differently than most. When Flores’s team ships their Marketing Super Agent on December 2nd, sales enablement won’t see the final product until after launch. Go-to-market leaders had to accept that sellers would be trained on evolving demos. It’s not just faster; it’s fundamentally different, requiring leaders to embrace imperfection and trust that failures can be fixed quickly. “You have to be okay with saying it’s not going to be perfect,” Flores notes, capturing the mindset shift required across entire organizations.

Perhaps his most valuable advice cuts through the noise: stop selling, start educating.

As big tech races to monetize AI, Flores believes vendors should focus on thought leadership first. Whether a CMO buys Treasure Data’s product or not, he’s willing to walk them through what an AI agent actually is. “There’s no such thing as a stupid question,” he emphasizes, acknowledging that AI evolves daily and no one, regardless of their title or technical background, can claim mastery.

For marketing leaders feeling pressure from boards and bosses to “do something with AI,” the permission to ask basic questions might be the most liberating takeaway. The technology is here to stay. The real competitive advantage goes to those humble enough to keep learning.