Prototyping & Design Marketing & Content

Jamey Gannon's AI use case

AI Creative Director at Brand Sprints

Uses mood boards, Midjourney style references, personalization codes, Nano Banana, Flora, and Figma to create consistent AI-generated brand imagery that can be packaged into reusable client-ready visual systems.

The problem

What was broken before AI

AI image tools can make beautiful individual images, but they often struggle with consistency. A creator might get one great result and then spend hours trying to make the next image feel like it belongs to the same world. For brand work, that is a real problem: the goal is not just a good picture, but a repeatable visual language that can survive across campaigns, thumbnails, social posts, and client assets.

What changed

What the use case made possible

Jamey built a workflow around visual direction instead of prompt length. She uses mood boards to define the vibe, individual style references to steer Midjourney, personalization codes to bring in her taste, and simple prompt shortcuts like publication names or camera references. When an image is close but has a bad object, weird detail, or low-resolution finish, she brings it into tools like Nano Banana or Flora for targeted edits and refinement. The final output can be packaged in Figma as a reusable brand kit.

Why this matters

Why this use case is worth studying

Jamey’s workflow makes AI art feel less mysterious. The creative control comes from choosing the right visual inputs, noticing what is pulling the image in the wrong direction, and making small adjustments. That is a very different posture from writing longer and longer prompts. It feels closer to art direction: gather the references, test the direction, remove what is throwing it off, and turn the winning pattern into a system other people can use.

Use this when

When this pattern applies

Use this pattern when you need AI images to feel like they belong to the same brand, campaign, or visual world. It works especially well when one good image is not enough and you need a repeatable style that can show up across social posts, ads, thumbnails, presentations, or client assets.

Exponential Builder analysis

Taste is the control layer.

Jamey’s workflow shows that consistency comes from curating the right visual inputs, then testing how each one changes the output. The builder skill is less about prompt cleverness and more about knowing which references deserve authority.

Diagnose the reference, then edit the system.

When an image drifts, the fix may be removing a mood-board image, cropping a composition reference, or swapping a style input instead of adding more words. That mindset turns AI image generation into a feedback loop you can actually manage.

Package the recipe, not just the asset.

The durable deliverable is the set of references, personalization codes, prompts, and examples that reproduce the look. For client work, that shifts AI imagery from a one-time production trick into a reusable brand system.

Who this is for

Best fit

Brand designers

Creative directors

Marketing teams

Content creators

Agencies building visual systems for clients

Founders creating consistent launch imagery

Anyone using AI images for more than one-off experiments

What to avoid

Mistakes and warnings

Where this pattern can go wrong if you copy it too literally.

Do not expect one good Midjourney image to automatically become a consistent brand system.

Avoid dumping too many mixed references into the model and hoping it understands the aesthetic.

Watch for one reference image skewing the whole output in an unwanted direction.

Do not solve every visual problem by making the prompt longer.

Fix near-miss images with targeted edits instead of regenerating from scratch forever.

Public workflow preview

The shape of the workflow

A high-level look at how the use case works, with the reusable pattern made clear.

Build the visual language first

Jamey starts with a mood board in Pinterest or Cosmos so the model has a clear sense of the world she wants to create.

Test references one at a time

Instead of feeding Midjourney a messy board all at once, she looks for the images that actually pull the output in the right direction.

Layer in personal taste

Midjourney personalization codes help the images feel more specific and less generic.

Use simple prompt shortcuts

Publication names, camera references, and plain human descriptions can guide the style without turning the prompt into a technical essay.

Refine the almost-right images

Tools like Nano Banana and Flora help fix details, replace objects, upscale images, or generate consistent self-portraits.

Package the system

The best references, codes, and prompts are delivered in Figma so the client can keep making images in the same aesthetic.

Copy the pattern

The reusable idea

Pattern in one sentence

Build the visual system before chasing the final image, so AI has a consistent world to draw from each time.

Reusable idea

Jamey’s approach is useful because it starts with taste before it starts with text. If you want consistent AI visuals, gather the world you want the model to understand: colors, textures, references, camera styles, compositions, and examples that feel close. Then test what each reference does instead of assuming the model understands your vibe. Over time, the repeatable asset is not just the image — it is the visual recipe that keeps producing images in the same family.

Steal this workflow

Use this mini-template for a reusable AI brand image kit:

Create a mood board in Pinterest or Cosmos with 20–40 images that capture the intended world: color, texture, lighting, composition, camera feel, and mood.

Run one simple Midjourney test using the full board to see the broad direction.

Test the strongest references one at a time as style references. Keep a note beside each: “helps,” “hurts,” or “use only for composition.”

Remove references that introduce unwanted color, contrast, saturation, objects, or mood.

Add the Midjourney personalization code you want to use as the taste layer.

Build 3–5 short base prompts using plain descriptions plus visual shorthand such as publication names, camera references, or scene details.

Use image references when layout matters. Crop them if they bring in distracting objects or details.

Send near-miss outputs to Nano Banana or Flora for specific repairs: object replacement, detail cleanup, upscaling, or self-portrait refinement.

Package the final system in Figma: mood board, approved style references, personalization code, base prompts, example outputs, and notes on what to avoid.

Suggested prompt

“Create [subject or scene] for [brand/use case] in the style of [publication, camera, or reference aesthetic], using the provided style references and this personalization code: [code]. Keep the visual language consistent with [colors/textures/mood/composition]. Include [specific scene details]. Avoid introducing [known unwanted details from prior tests].”

Field notes

Get new AI use cases in your inbox

A short weekly note on how real people are using AI to save time, make money, build tools, and run their lives.

Related use cases

Keep exploring nearby systems.

Browse all

Prototyping & DesignMarketing & Content+1

Adobe Firefly Custom Models

AI image-generation feature for training brand- or creator-specific visual models at Adobe

Brand-specific image generation workflow where creators and brands train private Firefly models on their own assets to generate visuals with consistent style, characters, and details.

Adobe FireflyFirefly Custom ModelsFirefly Boards+2

Use case

View use case

Marketing & ContentPersonal Productivity

Josh Belanger

Financial entrepreneur and newsletter writer at Belanger Trading

Finance newsletter editing workflow where human-written market thoughts are streamlined with ChatGPT, copyedited with Claude, and refined with custom GPTs to preserve trading-specific language.

ChatGPTClaudeCustom GPTs

Use case

View use case

Prototyping & DesignMarketing & Content+1

Jascha Kaykas-Wolff + Alexandra Roberts

Co-creators of Eleanore; Kaykas-Wolff is a former Mozilla CMO and Roberts is an artist at Eleanore

AI-assisted beverage brand buildout workflow spanning brand identity, market strategy, web design, vendor outreach, packaging, and operational planning.

ChatGPTCustom GPTsClaude+9

Use case

View use case