Staff shopping.
I’m fairly certain I don’t treat AI the way most people do. But then again—what else is new?
Without a background in formal technology or prompt engineering, I’ve had to muddle through on my own, trial-and-error style, with my fingers firmly crossed. But as I explored these tools, one thing became clear very quickly: expecting a single AI to do everything—from writing and image generation to acting as both teacher and student—is a recipe for mediocre results.
So, I changed my strategy. Instead of looking for an all-in-one platform, I started shopping around to build a dedicated digital studio team, assigning specific roles to whichever AI handles that task best.
The Roster
Copilot (The Creative Confidant): My go-to sounding board. Copilot excels as a conversational partner for brainstorming raw project ideas, clarifying intentions, and talking through creative directions before execution.
Gemini (The Technical Lead & Fact-Checker): Brilliant with technical details, context, and cross-referencing. When I need complex information vetted or structural logic sorted out, Gemini takes charge.
Meta AI (The Artistic Director): The missing piece that made everything click. Today, I officially promoted Meta AI to Artistic Director. Its visual capabilities bring immediate life to concepts, allowing ideas to translate visually in seconds—and for me, it all hinges on the images.
Bridging the Gap: Language Training
By far, the most challenging part of building this team was language training.
Copilot was simple enough—it is inherently geared toward fluid, conversational interactions right out of the box. Gemini, on the other hand, took real work. Before I managed to coax it out of its shell, Gemini's responses were highly technical, drawn-out, and dense.
I constantly had to balance how much technical detail I wanted to read against how much I needed it to understand me. Have you ever tried persuading a High Court barrister from a privileged, purely academic background to engage in a casual, down-to-earth conversation with an everyday person from the real world? That is exactly what it felt like. There had to be give-and-take on both sides. I had to learn how to frame my intent, while slowly coaching the AI to drop the textbook jargon and speak human.
Finding the Rhythm
Delegating roles across different AI platforms—and taking the time to train how we communicate with each other—has transformed what used to be a confusing, trial-and-error experiment into a smooth digital production workflow.
I might still be learning as I go, but at least now I’m managing a multi-talented team that speaks my language.