AI/ML
October 2, 2025
One year ago, on October 1, 2024, I made a decision that changed everything about the way I lead, create, and think about the future of video. Until then, I had only experimented at the edges of AI. One of my earliest steps was with HeyGen, where my team and I created avatars of colleagues to see how they might be used in marketing. It was an exciting glimpse of what was possible — a spark that opened the door to a much bigger journey. But I knew there was more to explore, and that’s when I decided to fully commit myself to learning what AI could really mean for creative video.
That day, I drew a line in the sand. I wasn’t going to treat AI as a side project anymore. I was going to dive in headfirst, immerse myself, and commit to learning as much as possible about how AI was reshaping creativity. If this was going to be the future of video, then I needed to know it inside and out.
The first thing I did was bathe myself in AI news — every single day. I searched for where the conversation was happening and who was documenting the story. At first, that led me to Curious Refuge, a YouTube channel that became my “gym” for AI. They gave me a place to see what was new while also digging into the archives to understand how this technology had evolved from SORA’s early experiments to Runway’s growing capabilities to Kling’s dramatic entrance onto the scene. Soon after, I added Theoretical Media to my daily lineup, a channel that went deeper into workflows and the nuts and bolts of how people were applying AI to creative work.
But it wasn’t enough to just watch. I joined the conversations wherever they were happening: Discord groups buzzing with experiments, Reddit threads dissecting every update, LinkedIn communities exploring professional applications, and Instagram creators posting wild tests. These weren’t just channels; they were living classrooms. It gave me the benefit of structured learning and the chaos of real-time discovery, and together they became my foundation.
And as I dug in, collaboration became just as important as the learning itself. I wasn’t on this journey alone — I had like-minded co-workers who were eager to explore alongside me, and we challenged, inspired, and pushed each other forward. I also found myself influencing how others thought about AI, both inside Relevate and across the broader industry, sharing what I was discovering and helping people imagine new ways AI could fit into their own work. That spirit of shared exploration turned what began as a personal pursuit into something much bigger — a collective momentum.
For the rest of 2024, I immersed myself. My nights and weekends were filled with videos, articles, and demos. I wasn’t just asking, “What can AI do?” I was asking, “How does this compare to the way we make video now? Where could it replace steps, where could it accelerate them, and where could it open entirely new doors?” I wasn’t chasing gimmicks. I was searching for workflows.
By December, opportunities to put all this learning into practice started to arrive. We began by applying AI to internal projects — sales meetings, highlight reels — safe spaces to test the tools under real timelines. Then, early in 2025, client work arrived. We brought AI into pitch concepts, using it to show possibilities that would have been impossible just weeks earlier. That was a critical turning point. Tinkering is fun, but it doesn’t get you very far. Real growth comes when the work is real — when there are deadlines, expectations, and consequences. That’s when AI stopped being theory and started becoming part of how we delivered.
By March, things started to click in an entirely new way. Tools were finally catching up to the creative needs we’d been dreaming about. One of the biggest breakthroughs was character consistency. Back in October, you couldn’t reliably generate the same face twice. If you’re a storyteller, that’s a dead end — you can’t string together a narrative when your characters change every frame. But by March, new features were emerging that made consistent characters possible from multiple angles. It seems like a small thing, but it was transformative. For the first time, we could stitch together true scenes, not just single clips.
Runway was the first platform to really pull this together. Already one of the biggest names in AI video, they added reference tooling, and suddenly consistent characters could move from stills to motion. That update kicked off what I call the “summer arms race.” Every week seemed to bring a new release: Midjourney, Flux Kontext, Kling, Nano-Banana, Seedream. Each one claimed to be the best, only to be overtaken days later. For creators, it was exhilarating — but it also demanded constant learning, testing, and adapting. The ground was shifting under our feet weekly.
And then came May. Google released Veo 3, and the entire industry tilted. Veo 1 and 2 had already been strong, but Veo 3 introduced something no one had pulled off at scale: dialogue. Characters didn’t just move anymore — they performed. Lip sync was sharp, gestures felt alive, and suddenly AI video wasn’t just a clever tool; it was cinematic. It was, without question, the most monumental moment of the year. Other tools scrambled to add audio — Kling, ByteDance — but none matched the polish of Veo 3. For many of us, it was the first glimpse of what AI video might truly become.
By summer, our entire team had shifted into an AI-first mindset. We weren’t asking, “Should we use AI?” anymore. We were asking, “Where should AI fit in?” Every project became an opportunity to test: could AI replace this step? Could it accelerate that one? Could it make the work more creative, not just more efficient?
And the results started to show. By fall, we were delivering entire projects — full training videos created entirely with AI. What once took months, we produced in weeks. Clients were astonished. And the numbers were clear: engagement with these AI-generated training tools was 15–25% higher than traditional approaches. It wasn’t just faster. It was better.
In recent weeks, my focus has shifted again. I’ve been chronicling workflows, documenting which tools belong where, and clarifying who on the team can own different portions of the process. It’s no longer about playing with the technology. It’s about systematizing it — turning it into a repeatable, scalable engine for creative work. Because one thing is certain: creative will change in the year ahead. The tools are moving too quickly, the efficiencies are too powerful, and the opportunities too compelling to ignore.
The biggest lesson of this first year is simple: you cannot put all your eggs in one basket. No single tool can do it all, and the landscape shifts too quickly for any platform to hold the crown for long. The smarter approach is to identify one or two staple tools you rely on and then build a broader toolbox around them. It’s like carpentry: sometimes you need a Phillips screwdriver, sometimes a flathead, sometimes a saw. Mastery comes not from owning tools, but from knowing which one to use, when, and why.
Looking ahead, Year Two will be about scaling and leading. At Relevate, we’re weaving AI into our core workflows, training our teams, and guiding clients through responsible, impactful adoption. But more than that, I hope my journey shows others what’s possible.
Dedicating yourself to AI isn’t about chasing a trend. It’s about giving yourself the chance to see your own field through entirely new eyes. It’s a little like being an archaeologist who uncovers a new branch of dinosaurs that no one even knew existed. Suddenly, the world you thought you understood is bigger, more complex, and far more exciting than you imagined. And in that moment, your passion for the work reignites.
For me, that commitment has already paid off. I’m more energized today than I was a year ago, because just when I think I’ve seen it all, another tool arrives and rewrites the map.
AI isn’t the future of creative work. It’s the present. And if the past year has taught me anything, it’s this: the real adventure begins when you decide to step into the unknown and keep digging.