Every campaign needs scroll-stopping creative, and generic AI image tools fail B2B marketers in two specific ways. The first is that the output is off-brand: generic, stock-looking images that ignore your palette, so they never quite fit the rest of your content. The second is worse, and it is the single most common AI brand disaster: the model hallucinates a garbled, wrong version of your company logo and drops it right into the creative.

We think image generation should read as your brand by default, and your logo should never be left to a model's memory. Here is how that works in DesignTech AI.

The real problem: off-brand output and a fabricated logo

Off-brand imagery is a slow leak. Each generic image is individually fine and collectively corrosive: a feed of visuals that do not share a palette or a style reads as inconsistent, and inconsistency undercuts a brand exactly where it is trying to look established.

The fabricated logo is a fast disaster. A model asked to render "our product ad" will often try to draw the logo from memory, producing a smeared, misspelled, subtly-wrong mark. Ship that once and you have done real brand damage. The fix is not "prompt more carefully." It is to put the real logo into the design yourself, and keep the model to the artwork.

Our take: your palette by default, your shape by choice

You describe the scene to the Studio in plain language, and its Generate an image capability makes it. You control what matters:

  • The shape. You pick the aspect ratio, square (1:1), portrait (4:5 or 9:16), or landscape (16:9 or 3:2), and the image is made at that shape. A route that cannot render it says so rather than quietly handing you a square.
  • The palette. Your Brand Kit's palette rides with every new image, so what comes back reads as your brand and not as stock.
  • References. Give it images from your Gallery to put in the picture, or point it at any Gallery item, a deck, a poster, a page, as the look the picture should be in. Up to sixteen images can go into one generation.
  • Edits. Give it a base image and say what to change, and it keeps everything the instruction does not name.

It can make up to four images from one description, and each one is filed in your Gallery as its own image asset, reusable in whatever you make next.

Two engines, chosen for the job

Images are made by GPT Image, from OpenAI, and FLUX.2, from Black Forest Labs. GPT Image is strong at multi-reference editing; FLUX.2 at generation, reference editing and cleaning text out of a picture. Both run under commercial terms that do not let the provider train on what you send.

Keep the logo out of the model

The rule we hold ourselves to: the model does not draw your logo. When the Studio designs a carousel, its artwork is generated under a standing instruction never to fabricate or reproduce a logo and never to add text of its own, and your real logo from the Brand Kit is set in the design as part of the layout. Do the same with the images you generate for your own creative: ask for the scene, and place the real mark yourself.

The artwork is the model's job. The logo is yours. Keep the two apart and the smeared-logo disaster cannot happen.

Describe the scene. Pick the shape. Let your palette ride with it. And put the real logo in the design, not in the prompt. That is AI imagery you can put in a campaign without a designer checking every frame for a mangled mark.

Related: put the imagery to work in a Social carousel, and turn it into motion.