AI Photography: Where It Fits in Commercial Work

A campaign image can now be made before a studio is booked, a location is scouted or a single light is switched on. That is the immediate appeal of AI photography: more visual possibilities, generated quickly. For brand and agency teams, though, the useful question is not whether AI can make an impressive picture. It is whether that picture can carry the idea, meet the brief and hold up wherever the campaign needs to live.

I see AI as another image-making tool, not a replacement for considered commercial photography. Used well, it can help a creative team test an idea, solve a specific production challenge or extend a visual system. Used carelessly, it can produce work that looks polished at first glance but lacks the detail, authorship and brand character that make people stop.

What AI photography actually means

The term covers a few different things, and they should not be treated as one.

Sometimes it describes an image made entirely with a generative model from a written prompt or reference material. Sometimes it means a photographed image that has been enhanced, rebuilt or expanded with AI tools. It can also mean using AI during development to create mood references, rough storyboards or proposed art direction before the final image is made.

Those uses have very different stakes. A loose visual reference in an internal presentation has a different job from a hero image for a national out-of-home campaign. A background extension for a digital crop is different again from generating a product that needs to be accurate down to its packaging finish.

The distinction matters because commercial images are not judged only on aesthetics. They need to be right: right for the product, right for the audience, right for the media placement and right for the brand. They must also be defensible when questions arise around usage, provenance, talent, intellectual property and representation.

Where AI photography can add genuine value

The strongest use of AI is often upstream, before the production decisions have hardened. An early concept can be difficult to explain with words alone. Generative imagery can give a team something tangible to react to: the angle of a scene, an atmosphere, a colour direction, the scale of a set or the energy of a casting approach.

That does not make the first generated image the answer. It gives the conversation a starting point. The value is in finding out what the idea wants to become, quickly enough to improve it.

Previsualisation that makes a brief clearer

For a complex food, lifestyle, architecture or product shoot, previsualisation can make a production more precise. We can test whether a high-key set feels too clinical, whether a surreal styling device has enough impact, or whether the composition leaves room for copy across a suite of assets.

This is particularly useful when several stakeholders need alignment before shoot day. Rather than asking everyone to imagine the same thing from a deck of references, the team can assess a proposed direction in visual terms. It often exposes issues early: the product is getting lost, the setting is too generic, the talent feels wrong for the audience, or the image is relying on an effect rather than an idea.

Building worlds that would be impractical to shoot

There are occasions when an imagined environment is the point. A beverage floating through an impossible landscape, an interior transformed beyond what a location allows, or a visual metaphor that cannot be built safely within budget can all suit a hybrid approach.

Even then, a photographed foundation is frequently valuable. Real light on a real pack, surface or person creates material with physical credibility. AI can then assist with an extension, a transformation or a composited setting. The result can have more presence than a fully generated image because it begins with details the eye recognises: true reflections, real material behaviour and a specific human expression.

Faster adaptation across channels

Campaigns increasingly need a family of images, not one perfect frame. There may be vertical social crops, wide digital placements, e-commerce details, point-of-sale work and print executions, each with different needs.

AI-assisted post-production can help explore alternative framing or background treatments, especially where an original shot has been composed with room to adapt. But it is not a licence to shoot without a plan. The most efficient approach is still to map the deliverables early and create the core imagery with those uses in mind. AI is better at helping an established visual system travel than rescuing a system that was never properly designed.

What it cannot replace

A convincing image is not necessarily a useful commercial image. AI can imitate the surface cues of expensive photography, but it does not understand why a particular hand position makes a product feel aspirational, why an architect cares about the line of a junction, or why a food stylist adjusts one small highlight for ten minutes.

Those decisions are the work. They come from observation, craft and a clear understanding of the intended audience.

Product truth and brand trust

If a client is selling a physical product, accuracy is rarely optional. Pack colours, proportions, labels, ingredients, finishes and functional details all matter. Generative tools can invent a near miss with complete confidence. A cap changes shape, type becomes unreadable, a glass behaves oddly, or a product gains a feature it does not have.

For some conceptual work, that looseness may be acceptable. For packaging, retail, food, beverage and design-led brands, it can become an expensive problem. A photograph gives us control over the product itself. We can light it, style it and retouch it with care while keeping faith with what customers will actually receive.

People, place and the particular

The same applies to people. A generated face can look believable, but believable is not always enough. Real talent brings timing, individuality and an emotional response that cannot be specified into existence. A real location carries its own texture and context. These qualities matter when a brand wants connection rather than visual wallpaper.

There is also a question of accountability. Who has approved the likeness? What material informed the generated output? Are there cultural, legal or reputational concerns in how a person, place or community is represented? The answers may vary by campaign, but they should be discussed before the work is commissioned, not after it has appeared in market.

The questions worth asking before using AI

AI photography works best when it is part of the brief, not a late instruction to make something cheaper or faster. Before choosing a route, I would want clarity on the campaign's central task. Is it to show a real product beautifully? Create a fictional world? Generate interest around an editorial idea? Produce a high volume of adaptable assets? Or establish a distinctive visual platform that can last beyond one launch?

Then come the practical questions: where will the imagery appear, how long will it be used, what degree of product fidelity is required, and who needs sign-off? A brand team should also understand the source and permissions behind any generated work, along with the terms governing its commercial use. These are not administrative details. They affect whether an asset can be used confidently at all.

Budget deserves an honest conversation too. AI can reduce costs in some areas, particularly during concept development or where a modest alteration avoids a reshoot. But polished results still require art direction, retouching, compositing, quality control and careful output for each channel. Cheap-looking work is not always cheap to make. More importantly, it can cost far more if it weakens a campaign's credibility.

A better model: use the right tool for the idea

The most interesting direction is not a contest between AI and photography. It is a more deliberate combination of imagination and evidence.

For some projects, the right answer will be a fully produced shoot: considered casting, carefully controlled light, strong styling and imagery with a physical authority that clients can use across every channel. For others, AI may sharpen a concept in pre-production, expand a world in post-production or create an element that would otherwise be out of reach. It depends on what the image needs to prove.

My role is to help make that call with the creative team. That means protecting what is essential to the idea, challenging assumptions where needed and building a production approach that serves the campaign rather than the novelty of the technology. Calm on set matters, but so does being exacting about every decision that reaches the final frame.

The images worth noticing will not be the ones that merely reveal how they were made. They will be the ones that give the brand clarity, character and a reason to be remembered.