A three-hour hands-on workshop for instructional designers and L&D professionals. You’ll spend time generating, refining, and fixing your own images, not just watching me do it.
The problem
Most AI images look like AI images. The people in them grin weirdly while avoiding eye contact with each other, the style changes between slide four and slide nine, and nothing quite matches the brand palette you were given. So you go back to the stock photo library and pick the one with the fewest fake smiles.
It isn’t the tools. The key is the process when a first result is wrong and you fix it. That iteration is impossible to fit into a one-hour conference session or webinar.
What’s different about this workshop
I’ve taught AI image generation at DevLearn, for ATD chapters, and in free webinars. Those sessions cover some of the same ideas this workshop. What I can’t do in sixty minutes is give you enough time to practice, especially the process of refining and iterating until you get something you can actually use.
This workshop is built the other way around. Six practice rounds, each one followed by a debrief where we look at what you actually got and figure out why. You pick a real course or topic at the start, and everything you generate is for that project.
Example: Cheesy AI image to realistic working image
Generic, vague prompts can lead to cheesy AI images. Prompts with more specific details can improve your results.
What you’ll do
- Round 1: Identify your project and goals.
- Round 2: Write your first prompt with the framework and check your results.
- Round 3: Take that image and make it better.
- Round 4: Make a second image that matches the first using a reference image.
- Round 5: Generate a small set of icons or illustrations with consistent style and colors.
- Round 6: Build a character, change their expression, and put them in a scene while maintaining character consistency.
What you’ll leave with
- A small set of images for a real project, made by you, in consistent styles.
- A prompt framework you can use in any tool and strategies for fixing prompts and iterating for better quality.
- A handout with the tools, the prompt patterns, the style vocabulary, plus additional tips for improving your results.
- The recording so you can review it and work through anything again.
Example: Create icons and illustrations in your brand colors
Learn how to generate custom icons and illustrations in matching styles and consistent colors.
Who this is for
Instructional designers, elearning developers, and L&D people who need images and don’t have a budget for custom illustration. It’s built for beginners, and it’s built for the people who’ve tried AI images and quietly gave up on them. If you’ve never generated an image before, you’ll be fine.
It’s probably not for you if you already have a working AI image process and just want to hear about new tools. That’s what my free webinars and blog posts are for.
What you’ll need
I strongly recommend a paid image generation tool for this workshop. The free tools are good, but we’ll generate a lot. Three hours of real practice could mean 30+ images, and every free tier has a daily cap you’ll probably hit somewhere in the middle of the workshop.
Here’s the good news: you may already have one. If your organization gives you ChatGPT, Microsoft Copilot, or Gemini, those will work fine for everything we’ll do. Check before you buy anything. If your organization gives you access to Claude, it won’t work for photorealistic image generation unless Claude is also connected to another tool like Flora.
If you don’t already have access to a paid AI image tool, here are my recommendations. Plan to pay for one month; you can cancel later if you’re not using it.
- My first choice is Gemini. The image model in Gemini (Nano Banana) is the best I’ve used for consistency and for editing images, which is most of what we practice.
- ChatGPT: Another solid choice for overall image generation, especially with the recent image model improvements.
- I use Midjourney for most of my blog posts. The $10/month plan gives you plenty of generations for what you’ll need in the workshop.
- Flora and Magnific (which you may know as Freepik) are both solid options too; they include Nano Banana, GPT Image, and other major image models. These are more powerful but have a little more of a learning curve. They’re great if you’re already doing some image generation and want to get more professional results. If you’re just beginning to work with image generation, Gemini is a safer starting point.
You can take this workshop on free tools, and I’ll show you which ones stretch furthest. Just know you’ll likely be switching tools partway through, and I’ll walk you through that when it happens.
Examples: Consistent character across multiple scenes
Using reference images, clear prompts, and some iteration, you can generate images of the same character in multiple scenes, poses, and shots.
Details
- From First Try to Finished: Hands-On AI Image Generation
- Wednesday, October 14, 2026, 1:00 to 4:00 PM ET. Three hours, live on Zoom, with a break. Check your time zone.
- Limited to 35 people so there’s time to look at everyone’s work.
- Set up your tool before the live session. I’ll send instructions when you register.
- Recorded. Everyone registered gets the recording and the handout.
- $199
About me
I’m Christy Tucker. I create elearning and branching scenarios, which means I need a constant supply of images of specific people doing specific actions in realistic work settings. That mean stock libraries never quite had what I needed. I’m not an AI researcher. I’m an L&D practitioner who has generated a very large number of images and learned what works by getting it wrong a lot.