If you’ve ever tried adding more objects to an already-generated image inside Sheets, you’ve probably seen it happen. Details start to degrade. Text becomes less sharp. Materials lose realism. Scale starts to feel slightly off. The problem usually isn’t the model. It’s how we’re editing the image. When you repeatedly edit the same generated image in Sheets, each new generation compresses and reinterprets the last one. Over time, clarity drops. Fine detail gets re-rendered again and again. And the image slowly loses precision. It’s not a capability issue. It’s a workflow issue. So we redesigned the workflow. This method works for interiors, beauty setups, fashion styling, food scenes, retail displays, or any situation where multiple objects need to exist in the same frame at correct scale.

What Each Role Actually Does in a GenAI Team From strategy to scale, these are the roles that make it work. Who owns what. What good looks like. What skills matter. Roles, responsibilities, and the skillsets that make the difference.

Almost six months ago, I wrote about how to set up your creative team for AI. And I already have to refresh it. That’s how fast AI is moving. One of the biggest shifts since then has been the release of Workflows, a nodal-based UI that lets you visually map insights to ideas to outputs (including revisions). Instead of bouncing between chats and sheets, you now connect everything together: context, attachments, logic, references. You can literally see why you’re doing what you’re doing. An insight flows into a persona. A persona flows into a concept. A concept flows into images, videos, and copy. Nothing sits in isolation. Nothing loses its origin story. And when you can see the logic, you start designing teams differently. “When the system becomes visible, new roles emerge.”