This session offered a hands-on look at how image prompting works and why it plays such an important role in creative storytelling. Led by Jason Baxter, Senior Cloud Solutions Architect at Microsoft, the session focused on using AI to iteratively refine prompts and generate high-quality visual content.

The Goal and the Guardrails Why Image Prompting Matters 

The goal of the session was simple but effective: create a teddy bear image through iterative refinement. Along the way, Jason emphasized the importance of working within clear guardrails to ensure responsible and brand-safe use of AI. 

These guardrails included: 

• Using only original assets uploaded by the user 
• Avoiding imitation of living artists 
• Ensuring all outputs remain brand-safe 

Why Image Prompting Matters 

Jason highlighted that image prompting unlocks creativity and visual storytelling, enabling more engaging and effective communication. Unlike text-based prompting, image prompting introduces a new way to express ideas visually. 

Image prompting differs from regular prompting in several key ways. The output focuses on images, illustrations, and edits rather than text. Prompts rely on different ingredients, such as subject, style, lighting, and composition, and use more descriptive, sensory, and artistic language. Iteration becomes central to the process, with refinements made to style, mood, and layout. Users also gain greater control over elements like art style, colour, and setting, while evaluation shifts toward visual appeal, brand fit, and creativity. 

The Iterative Image Prompting Process 

Jason walked through a clear, step-by-step process for image prompting. 

The process began with a clear, descriptive base prompt, with specificity improving results. 

From there, customization layers were added incrementally using the Transform option. This allowed changes such as: 

• Adding accessories 
• Refining the setting 
• Exploring different poses 

Jason demonstrated the clear difference between outputs generated from a basic prompt versus a hyper-detailed prompt, showing how each refinement improved the final result. 

The session also showed how AI itself can be used to optimize prompts, helping users achieve better outputs with less trial and error. 

More Than Just Images 

Beyond static images, Jason demonstrated that the same prompting techniques can be used to create a wide range of content, including: 

• Videos 
• Infographics 
• Other visual formats 

Our Takeaways 

This session reinforced how powerful iterative refinement can be when working with AI-generated visuals. 

Key takeaways included: 

• Iterative prompt refinement has a significant impact on image quality 
• Image prompting is accessible and approachable with AI 

A big thank you to Jason Baxter for showing how structured experimentation and thoughtful prompting can unlock creativity and produce compelling visual content.