← All Articles
AI & Automation7 November 20264 min read

Should You Use AI Images In Your Marketing

Yes for ideas, concepts and scenes that are clearly not photos of your work. Never for your team, your finished jobs or your customers, because a fake photo of proof is a false claim.

Use AI images for ideas, concepts and scenes that are clearly illustrations, and never for anything a buyer reads as proof: your team, your finished work, your customers or your reviews. A generated photo of a job you never did is a false claim with good lighting.

We generate images at Ignis every week, for real client work, so I am not telling you to stay away from the tools. One rule sits above every prompt we write, and it is real assets first.

Where AI images earn their place

Concepts before a shoot. A generated frame shows the videographer and the client what you are going for, and nobody has to imagine it. Nobody outside the room ever sees it.

Ad concepts for testing. You can try five visual ideas against each other before paying for a shoot, then go and film the one that wins with real people. How that fits a proper testing process is in AI-powered ad creative.

Backgrounds and textures. An abstract background behind a headline claims nothing about your business.

Things that do not exist yet. A developer selling off the plan can turn a real architectural render into an image of the finished building. Label it as an artist's impression, as developers have done for decades.

Where only real photos will do

Your team. If the person in the hard hat does not work for you, you have just invented a staff member.

Your projects. Buyers use those photos to decide if you can do the job.

Your customers, reviews and before-and-after shots. These exist to prove something happened, so a generated one is simply a lie.

Your logo. Use the real file every time, because a model will redraw it slightly wrong.

The Australian Consumer Law bans misleading conduct in trade, and it does not care which tool made the picture. A photo of a project you never built tells the buyer something untrue about your experience. So real photos need to come first, and the way to get good ones is in how to photograph your own work.

What buyers respond to

Real beats polished more often than people expect. A real person's photo with a caption that has some feeling in it beats a designed render that looks corporate. I would take a dozen honest phone photos over one perfect image of nothing.

Part of this is that everyone can make the perfect image now, so it has stopped signalling anything. I wrote more about that shift in how AI is killing generic marketing.

The rules we run at Ignis

These sit in our process so that nobody has to remember them on a busy day.

Real references first. We check the client's photo library before any prompt, and we feed real photos in as references so the output starts from the real thing.

Logos come from the file. We place the logo in afterwards, and it is never generated.

Words come from code. Image models still warp and misspell text, so headlines get added in the design layer.

Brand colour goes on after. Typing a hex code into a prompt does not hold, so colour gets corrected afterwards to match the brand exactly.

No real references means a person signs off. An image made from a text prompt alone never publishes itself, and the client approves it first.

The platforms label it anyway

Meta adds an AI info label to ads made with its own AI tools or with outside tools that it detects, and it has been expanding automated detection. TikTok requires creators to label realistic AI content. So plan on the audience finding out, and never try to strip the signal to get around it.

The bigger risk is the expert in your audience. A builder spots a scaffold that would never pass in about a second, and an electrician sees the wrong switchboard just as fast. Once they catch one fake detail, they start doubting the real work too. Check the hands, the signs, the safety gear and the tools in every image before it goes out.

The test before you post

Ask one question of every image. If a buyer rang tomorrow and asked whether it was real, would you be happy giving them the answer?

If you would, post it. If the honest answer makes you squirm, go and take the photo. Video follows the same logic, as covered in AI video creation for marketing.

Write that line down once for whoever makes your creative, and you keep the speed AI gives you on ideas while your proof stays real, which is the part that closes the sale.

David Eid

David Eid

Marketing Strategist · Founder of Ignis

Marketing strategist based in Sydney, Australia. Founder of Ignis - premium marketing that scales businesses. Our average client generates $3M+/year and 1M+ views/month.

AIImagesCreative
Share

Get weekly marketing insights

No fluff. No spam. Strategies from someone who has done it. Delivered to your inbox every week.

Subscribe