I learned the pen tool in 2019. Got fast at it. Could knock out a clean mask in 8-10 minutes on a straightforward product, maybe 20 on something complex like jewelry or a plant.
For six years, that was my workflow. Shoot, import, pen tool, refine edge, output. Repeat 200 times per client project. My wrist hated me.
Last year a photographer friend showed me his workflow. Same camera, same lighting setup, same products. But instead of Photoshop, he was running everything through an AI tool and only opening Photoshop for the 10-15% that needed manual fixes.
I thought it was lazy. Then I tried it. Then I couldn't go back.
Here's the honest breakdown of what changed and what didn't.
The Three Real Options
Skip the 20-tool comparison articles. In practice, you're choosing between:
AI tools — Upload image, get result in 3-5 seconds. Works great on most products. Struggles on some.
Photoshop manual masking — Pen tool or Select Subject + Refine Edge. Full control. Takes 8-25 minutes per image depending on complexity.
Outsourced retouching — Send to a service, get results in 12-48 hours. Costs $1-5 per image. Quality varies wildly.
Most sellers I know use some combination. The question is what ratio.
Where AI Wins (And It's Not Close)
Speed on Easy Products
A supplement bottle on a white background. A phone case on a gray surface. A pair of sneakers on a sweep.
These take me 8 minutes in Photoshop. The AI does them in 3 seconds. The quality difference? I can't tell them apart at listing resolution. Maybe if I zoom to 500% I'll find a pixel that's different. At the size customers actually see the image, they're identical.
When you're processing 50-100 images for a catalog refresh, that time difference is the difference between a full day and a lunch break.
Consistency Across a Batch
Here's something nobody talks about: when I manually mask 50 images in a row, the quality drifts. Image #1 gets my full attention. Image #37 gets my tired-eyes-at-11pm attention. The edges are slightly rougher, the feathering is slightly inconsistent.
AI doesn't get tired. Image #1 and image #50 get the same processing. For a product line where visual consistency matters (and it always matters), that's a real advantage.
The "Good Enough" Threshold
I used to agonize over every pixel. Then I realized: the customer is looking at a 500px thumbnail on their phone. They're not downloading the full-res image and inspecting the edges at 400% zoom.
For marketplace listings, "good enough" is genuinely good enough. And AI hits "good enough" on about 80% of products without any human intervention.
Where Photoshop Still Wins
Fine Jewelry
Chains. Thin prongs. Transparent stones. Reflective bands.
I've tested every AI tool I can find on a delicate gold chain necklace. None of them get it right. The links are too small, too reflective, and too close together. The AI either merges them into a solid line or loses individual links.
For jewelry, I still use the pen tool. But only for the hero shots — the main listing image and maybe one detail shot. The lifestyle images and secondary angles go through AI.
Hair and Fur
If you sell wigs, pet products with fur models, or anything with fine hair against a complex background — AI is getting better but it's not there yet. The edges are either too sharp (looks cut out) or too soft (looks blurry).
Photoshop's Refine Edge with the hair-specific brush is still the best tool for this. It's slow, but the result is noticeably better.
Creative Compositing
When I need to place a product into a specific scene with specific lighting and shadows that match — that's still Photoshop territory. AI can remove backgrounds and AI can generate scenes, but matching the lighting direction and shadow angle between a real product and a generated background still requires human judgment.
The Hybrid Workflow That Actually Works
Here's what I do now for a typical 50-SKU project:
Step 1: Shoot everything. Focus on product lighting. Don't worry about the background at all. I shoot on gray now because it's easier on my eyes than white, and it doesn't matter since the background is getting removed anyway.
Step 2: Batch through AI. All 50 products, all angles. Takes about 5 minutes total. I use pic1.ai because it handles the centering and platform sizing automatically, but the removal step is similar across tools.
Step 3: Quick review. Flip through all results. Flag anything that needs manual work. Usually 5-8 images out of 50.
Step 4: Photoshop the flagged ones. This is where my pen tool skills still earn their keep. But instead of 50 images × 15 minutes = 12.5 hours, it's 7 images × 15 minutes = under 2 hours.
Step 5: Final check and export. Platform-specific sizing, format conversion, upload.
Total time: about 3 hours for 50 products. Used to be 2-3 days.
The Money Conversation
I charge clients per project, not per hour. So when AI cut my editing time by 75%, I didn't lose income — I gained capacity. Same project fee, finished in one day instead of three. Take on more clients, or take Wednesdays off. I chose both.
For sellers doing their own photos: the math is even simpler. Your time has a value. If you're spending 15 minutes per image in Photoshop and you have 200 images to process, that's 50 hours. At any reasonable hourly rate, the AI tool pays for itself on the first project.
The Honest Take
Photoshop isn't going anywhere. I still open it every day. But I open it for creative work now — compositing, color grading, retouching. The mechanical work of cutting products out of backgrounds? That's AI's job.
If you're still pen-tooling every single product image, you're not being thorough. You're being stubborn. I know because I was that person for longer than I should've been.
Start with the easy products. See the results. Then gradually let the AI handle more. You'll find your own comfort level. Mine turned out to be about 85% AI, 15% manual — and that ratio keeps shifting toward AI as the tools improve.
Try it on your next batch. pic1.ai is free to test. Worst case, you waste 5 minutes. Best case, you get two days of your life back.
If you're ready to make the switch, here's my actual background removal process for 50+ products per month. And for the common pitfalls, read about the 7 background removal mistakes that quietly killed my conversion rate.