AI in Post-Production With the Photographer in the Driver's Seat

There are two very different ways AI can show up in a photographer’s workflow. One version makes decisions for you and asks you to accept the output. The other gives you automation when you want it and gets out of the way when you do not. Imagen is built around the second version, and the redesigned Studio experience makes that stance more visible than ever: you are still the one steering, from the first culling pass to the final export.

Here is what that actually looks like across a real project.

Two ways AI can show up in your workflow

Think about the tools you already use outside of photography. Some quietly take over a task and hide the mechanics from you entirely. Others hand you a strong starting point and expect you to adjust it. Photography is not a field where the first approach works well, because taste, client relationships, and a shoot’s specific story are things a photographer understands better than any model. Imagen leans deliberately toward the second approach: fast, capable defaults, with manual control sitting right next to them at every stage.

The failure mode of the first approach is familiar to anyone who has used a fully automated tool outside photography: a result that is technically consistent but occasionally, confidently wrong in a way that is hard to fix after the fact, because the system never exposed the decision it made. Applied to a wedding gallery, that might mean a skin tone correction applied uniformly across a couple with very different complexions, or a crop that technically follows a rule but ignores what is actually happening in the frame. Photographers do not have the luxury of shrugging that off, a client is going to look at every single photo.

Automation on demand, not automation by default

The clearest example is the project structure itself. Culling, editing, review, and export now live in a single project, in whatever order you choose to work through them. If you want AI culling, it is there. If you already know your selects and would rather cull manually, you can skip AI culling entirely at no cost and go straight to editing. Nothing about the structure forces a specific sequence on you, which matters more than it sounds, because different shoots genuinely call for different approaches.

Where letting go of a step actually helps

None of this is about resisting automation for its own sake. Letting the culling model do a first pass on five hundred wedding frames, or letting the editing model apply your Personal AI Profile across a shoot, saves hours that would otherwise go into repetitive work with no creative upside. Edited photos now arrive in batches by timestamp as they finish, so even the waiting that automation requires is broken into smaller, more useful pieces instead of one long block. The time you get back from automation is time you can spend on the parts of the job that actually need a human eye.

None of that requires you to supervise the automation closely either. You can start a culling pass on one shoot, move to reviewing an already-edited batch from a different one, and come back later to find the first shoot’s culling finished and waiting, without babysitting a progress bar in between.

Where your judgment has to lead

The flip side is just as important. A built-in editor, currently in beta, sits right next to the automated tools, with manual sliders and a histogram, crop and straighten, presets you can import from Lightroom or build yourself, and versions so you can branch a photo to try something different without losing the original. Quality here is comparable to Lightroom, though not every AI tool is available inside the editor in this first release, so treat it as a strong early version rather than a finished one-to-one swap. When a specific portrait needs a hand-tuned edit that no automated pass would guess correctly, that control is right there, not in a separate application you have to remember to open.

Culling a shoot inside the Imagen project view

An assistant that answers to you

The assistant built into the Studio experience follows the same philosophy. You describe what you want in plain language, find every five-star frame, find the shots from a specific moment, and it works through the gallery on your behalf. It does not act unless you ask it to, and it does not make creative calls about what counts as a good photo. It is also upfront about its limits: it knows what is in a photo but not exactly where within the frame, so it cannot place a mask on a specific spot yet, it does not generate images, it takes no voice input, and very subjective requests can still be a stretch. Within those boundaries, it removes searching and sorting time without ever taking the decision away from you.

You can also apply more than one AI Profile to a chosen set of photos inside the same project, for example running a distinct black-and-white treatment on just a handful of reception frames while the rest of the gallery keeps its standard edit. That kind of selective control, choosing exactly which photos get which treatment, is a small detail that says a lot about the overall approach: broad automation available everywhere, precise control available exactly where you want it.

A manual editor for the moments that need one

Even with strong automated defaults, every photographer has a shot that needs a specific human decision: a difficult skin tone, an unusual lighting situation, a client request that does not fit a preset. That is what the built-in editor is for. Because it lives inside the same project as everything else, reaching for manual control does not mean exporting somewhere else and losing context. You stay in one place, moving between automated and manual work as the photo demands it.

That same pattern holds for the more unusual cases too, a portrait session where one client requests black-and-white for the whole gallery instead of just a few frames, or a product shoot where three hero images need pixel-level attention no automated profile would attempt. The editor is there specifically for those moments, not as a fallback for when automation fails, but as a standing option you reach for by choice.

A concrete example: the same shoot, two approaches

Take a corporate headshot day, forty employees back to back. Automation on demand means you can let the editing model apply your Personal AI Profile across the entire batch, handling exposure, skin tone, and color consistency the same way for every person, which is exactly the kind of repetitive, high-volume work automation is good at. You did not need to touch a single slider to get there.

Now say three of those forty people need something different: one asked for a slightly warmer tone to match a brand guideline, one has a stray reflection in their glasses only a person would catch, one simply looks better with a different crop than the batch default. That is where the built-in editor and its manual sliders take over, on exactly those three photos, without reprocessing the other thirty-seven or forcing you into a separate application to make three small adjustments. Automation handled the volume, you handled the judgment calls, and neither one got in the other’s way.

Pricing built around your choices

The control-on-demand idea extends to pricing too, for pay-as-you-go photographers. You can test a tool, compare looks, or try a profile on a sample of frames at no cost, and pay only once you export, at the same per-tool price as before. Limitless plans are unaffected, and new accounts can explore properly, for free, before paying for anything on export. Frames a tool has nothing to act on, such as a retouch tool finding no face in a landscape shot, are not charged either. The system rewards exploring your options rather than penalizing it, which is really the same driver’s seat idea applied to your budget instead of your editing choices. For a closer look at how these pieces fit into a real shoot, our valuable tips library is worth a visit.

What’s new in Imagen: Imagen is now an intelligent, agentic Studio for professional photographers, built so you stay in charge from the first cull to the final export. The assistant finds photos and acts on them when you ask in plain language, connected to Imagen’s tools and with you steering every step, and you can skip AI culling and cull manually for free. The Studio is rolling out gradually as a beta.

Frequently asked questions

Does Imagen make editing decisions without my input?

No. Automated tools give you a strong starting point, but every decision about what ships to a client, from culling to final export, stays with you.

Can I turn off AI culling and do it manually?

Yes. You can skip AI culling when you set up a project and cull manually at no cost.

Can the assistant edit photos on its own?

It can find and help you act on photos based on plain-language requests, but it does not generate images or make independent creative decisions.

Is the built-in editor meant to replace my current editing tools?

Not yet, and not entirely. It is a capable beta with real manual control, and Lightroom remains part of the ecosystem for anyone who wants to finish there.

Want automation and control on your own terms? Start a project in Imagen’s new Studio at imagen-ai.com and see how it feels in your own workflow.