Creative with an idea in it, made at scale.

Campaign assets produced with AI, starting from a concepting system so every ad carries a marketing angle and not just a picture.

Done as a one-off project, or month to month within ongoing management.

AI has made it cheap to produce an ad, and just as cheap to produce a forgettable one. Most AI creative is a tidy image with a headline on it. Nothing in it says why this product, for this person, in this placement.

I start somewhere else. Before anything is generated, a concepting system built on a library of real campaigns works out the angle: the persuasion mechanic, the story frame and the format that suit the audience and where the ad will run. The free campaign concept tool on this site is a small version of it.

Production comes after that. The chosen concepts are turned into the full set of assets a campaign needs, across funnel stages, personas and ad shapes, using language models, image and video generation and templated rendering. Because the idea is fixed first, a hundred versions still say the same thing.

None of it is made in isolation. The assets are built to fit the campaign plan, the landing page and the sales journey, and they’re named and tracked so each concept can be measured, compared and improved.

This is for you if

  • You need far more ad variations than your team or budget can make by hand.
  • Your AI-made ads look fine and say nothing.
  • You have several audiences or funnel stages and show them all the same ad.
  • You’re launching a campaign and need image, video and supporting documents that match.
  • You’re an agency that needs production capacity behind a creative idea.

By funnel stage

Awareness, prospecting, content, sales and remarketing each get their own version of the idea. Someone who has never heard of you needs a different ad from someone with a full cart.

By persona

Each target audience gets the angle and the proof that matter to them, in place of one message for everyone.

By placement

Square, portrait and vertical images, display sizes and short video, each built for where it runs and how it’s seen there.

Beyond the ad

Landing page sections, whitepapers and PDF guides that carry the same idea through to the enquiry or the sale.

How the work runs.

  1. Brief

    The product, the audiences, the goal and the plan the creative has to serve.

  2. Audit

    Your current ads sorted by funnel stage, to show which stages have nothing made for them and which are the same ad with a different button.

  3. Concept

    Directions from the concepting system: angle, mechanic, frame and format, with the reasoning. You choose which go forward.

  4. Matrix

    A grid of what’s needed, stage by persona by placement, so nothing is made twice and nothing is missing.

  5. Produce

    Copy, images and video generated and rendered through templates, then checked against your brand rules by a person.

  6. Launch

    Assets named and tagged so each concept, stage and persona can be told apart in the results.

  7. Optimise

    Weak concepts dropped and strong ones extended, based on what the tracking shows.

Deliverables

What you end up with

  • An audit of existing creative by funnel stage
  • Concept directions with the reasoning
  • Asset matrix by stage, persona and placement
  • Image ads in square, portrait, vertical and display sizes
  • Short video ads
  • Whitepapers and PDF guides where the plan calls for them
  • Naming and tracking so every concept can be measured
  • Templates and prompts your team can reuse

Where AI comes in

The production line, not the idea

This is where AI does the most work. Language models draft the copy and the concept rationale, image and video models produce the visuals, and a rendering service turns templates into every size. The brand rules are written down, a person chooses the concept, and nothing goes live without someone approving it.

How I use AI

Fair things to ask.

Something else? Ask me directly.

Is this just AI-generated images?

No. The concept comes first: the angle, the persuasion mechanic and the format are chosen before anything is generated. AI is how the assets get made in volume, not where the idea comes from.

Which AI tools do you use?

It depends on the job. Commonly a language model such as Claude or ChatGPT for copy and concepts, a generation tool such as Higgsfield for images and video, and a rendering service such as Bannerbear to produce every size from a template. The tools change often. The process around them doesn’t.

Will it look like our brand?

Yes. Your fonts, colours, logo rules and tone are written into the templates and the instructions, and a person checks the output against them before it’s used.

Can you work with our designer or creative agency?

Yes. I can produce from their idea and brand system, or hand them concepts and a matrix to produce from.

How is it measured?

Every asset is named for its concept, stage, persona and format, so results can be read by idea and not just ad by ad. That shows which angle is working, which is what you need to know before making more.

Is it a project or ongoing?

Either. A campaign needs its set of assets up front. Ongoing work adds fresh concepts and versions each month as the results come in.

Something driving you mad?

Tell me what’s going on. I’ll come back with questions or a scoped proposal, and I’ll say so if I’m not the right person for it.