AI Culling and Editing Tools: How to Judge Them

Judge an AI culling or editing tool by running it on a job you have already finished and comparing its output to your own. Measure the time it saves, count the mistakes you had to fix and check what it does with your clients' images. A tool that saves hours without changing your style, and without sending files somewhere your contract does not allow, is worth paying for. One that does not should wait.

Prices and features change often, so check each vendor's site rather than trusting a list in a blog post.

What they do

Most tools fall into a few groups.

Culling tools read each image and group or rank them. Common abilities:

• Spotting blur and missed focus
• Finding closed eyes and awkward expressions
• Grouping near-duplicates and picking one
• Ranking by estimated quality
• Sometimes recognising faces to find the couple or key people

Editing tools apply adjustments learned from your past edits or a preset:

• Matching exposure and white balance across a set
• Applying a look to groups of images
• Straightening, cropping suggestions
• Skin retouching and object removal, in some tools

Delivery tools build galleries, albums or slideshows with automatic layouts.

Understand how each one learns. Some use general models; others learn from your own catalogue. The second kind needs a good set of your past edits to start.

Where they help

Think about the parts of your workflow that are repetitive and low-stakes.

• First-pass culling. A wedding of 4,000 frames may include hundreds of obvious rejects. Removing those automatically can save time before you make real choices.
• Duplicates. Bursts of ten similar frames are easy for a tool to group.
• Consistent base edit. If your style is steady, a learned profile gets you most of the way, so you spend time on the pictures that need care.
• Big volumes. School events, sports days and conferences with thousands of similar images benefit most.
• Rush deadlines. A same-day preview can use a quick automatic cull.

Example, only for illustration: if you spend four hours culling and a tool brings it to two hours of review, the saving is two hours a job. With 40 jobs a year, that is 80 hours. Your numbers will differ, and you should measure them.

For the wider process, see your editing workflow and turnaround time for how time saved translates into delivery promises.

Where to stay hands-on

Some decisions are about meaning, not pixels.

• The story. A tool does not know that the blurry picture is the only one of a grandmother's blessing.
• Expressions and moments. A technically perfect frame can miss the moment. A technically flawed one can be the best.
• Final selection. Choose the delivered set yourself. That is what clients are paying you for.
• Faces and skin. Over-retouching looks fake and can offend. Check each portrait.
• Colour in tricky light. Mixed lighting, stage lights and coloured walls fool automatic white balance.
• Sensitive work. For images that need privacy, such as boudoir, think carefully about cloud processing.
• Deleting. Do not let a tool delete files. Move rejects to a folder and remove them after delivery. This matters for your archive policy.

A rule many photographers use: let the tool propose, and you decide.

Testing on a past job

Never trial a tool on a live job. Do this instead.

1. Pick a finished job you know well, with a mix of lighting and moments.
2. Time your own cull from memory, or use your earlier notes.
3. Run the tool on the original files, in a copy of the folder.
4. Compare: how many of your selects did it keep, how many of its picks would you reject, what did it miss?
5. Note the time: setup, processing, your review and fixing.
6. Check edited output at 100% on a few key frames: skin, skies, whites.
7. Repeat on a second job of a different type, such as a dark venue.
8. Read the vendor's privacy and licence terms, and check where your files go.

Score it simply: time saved, mistakes fixed and trust. If the tool needs you to double-check everything, it saves little. Keep your test notes; the second run in a year tells you if tools have improved.

Telling clients

Be open about the parts that matter to clients: where their images go, who sees them and that you choose the final set.

• A line in your contract or FAQ can say: "We use software to help sort and prepare images. All final selection and editing decisions are made by our team."
• If you use cloud tools, say that images may be processed on a third-party service and that you have checked its terms. Ask a local adviser about data rules in your area.
• If a client asks, answer straightforwardly.
• Do not advertise AI as a feature. Clients care about delivery date and quality.

Also be clear in your own team. If an editor uses tools, agree which, and check they follow your rules. See the earlier article on gallery delivery options for how finished images reach the client, and the thank-you gifts post for what to do after delivery.

How ClickBook helps

ClickBook does not cull or edit pictures. Its edit queue tracks the stages around them: editing, studio review, client review and delivered, with versions, comments and approvals, so whoever edits, with or without AI, the status is visible. See the demo to look at the queue.

Test any tool on your own old work first, and keep the final eye for yourself.

Frequently asked questions

Can AI cull a wedding for me?

It can do a first pass: flag blur, closed eyes and near-duplicates. You still need to choose the images that tell the story. Treat its output as a suggestion.

Do I have to tell clients I use AI tools?

There is no universal rule. Being open about your process is a good habit, and some clients and contracts may ask. Check any local rules on data and privacy with an adviser.

Are AI tools safe for client images?

It depends on the tool. Check whether processing runs on your computer or in the cloud, what the vendor says about storing and training on images and whether that fits your contract.

How do I know if a tool is worth the price?

Run it on a past job and measure time saved and the changes you had to make. If you spend as long fixing its choices as you would culling, skip it.