Mastering AI Imagery: Jamey Gannon's Midjourney Techniques

Every Monday, the latest AI tips and tricks come alive through Claire Vo's podcast. The focus is always on the tangible, no grand theories, just practical, actionable takeaways. This time, AI creative director Jamey Gannon steps into the spotlight to discuss her expertise with Midjourney, a tool she’s mastered over countless hours to produce high-quality brand imagery.

A New Approach to AI Imagery

Gannon's method breaks the mold. Instead of relying heavily on text prompts, which many find cumbersome, she leans on mood boards, style references, and personalization codes. Her philosophy is straightforward: “Avoid prompting at all costs.” This isn’t just advice, it’s a strategy rooted in experience.

Is it time to rethink how we interact with AI? While complex prompts can sometimes seem like the best path, Gannon argues that image references and style guides might actually yield better results. She highlights the power of style references (SREFs), which often outperform traditional mood boards.

From Inspiration to Implementation

Gannon’s workflow is about turning loose inspiration into a scalable visual system. She warns against the pitfalls of averaging out styles seen in traditional mood boards. Instead, she suggests using precise style references to guide the AI more effectively.

One standout technique is the use of personalization codes. By rating images carefully during this process, Gannon can consistently produce her preferred aesthetic. It’s about creating a unique visual language that AI can replicate.

Editorial and camera references also serve as shortcuts. Gannon’s approach to bypassing verbose prompts is by simply referencing well-known publications or camera models, trusting that Midjourney can fill in the gaps. This is a testament to understanding the tool’s capabilities.

Adapting to AI’s Perspective

Gannon stresses the importance of seeing things from the AI’s perspective. When something isn’t working, her advice is to remove the problematic element entirely. This might mean excising an overly dominant reference image.

She also discusses the strategy of using different tools for different tasks. Midjourney handles initial creation, Nano Banana refines specific elements, and Flora manages the branding package delivery.

Ultimately, Gannon’s approach challenges conventional methods. Are we too locked into traditional prompting methods? Her insights suggest that a shift towards visual communication with AI could unlock new creative potential.