7 Checks Before Putting AI Product Photos into a Print Catalog

Before AI product photos go into a catalog, look at seven things: product proportions, shadow direction, material realism, edge artifacts, background consistency, whether colors are over-beautified, and the size relationships of multiple products on the same page. Minds Printing (MS, mid-to-high-end fully custom commercial printing) breaks those seven checks into a three-stage proofing gate: ① page-level believability, ② printability, ③ no misleading selling

I often see clients drop AI-cleaned product photos straight into a catalog. They look clean on screen, but once the proof comes back the edges look cut-out, the shadows fight each other, and the mug on the same spread somehow looks bigger than the thermos

E-commerce images can lean on brightness, sharpness, and background mood to pull a click. A printed catalog has to make the buyer believe 'this is roughly what I'll get in hand.'

・① Page-level believability: product size, angle, shadows, and how they line up with other items on the same spread have to make sense. When 10 SKUs share one page, the reference baseline matters twice as much

・② Printability: cut-out edges, resolution, dark-area gradation, and highlight detail have to survive being enlarged on paper. An A4 page is a lot more honest than a phone screen

・③ No misleading selling: color, material, sheen, and texture can't be polished until they don't match the real product. The four categories that blow up fastest are fabric, wood grain, metal, and food packaging

In a catalog, 'looks good' only ranks second. 'Looks believable' ranks first

I've said this at the press desk more times than I can count, because a reprint doesn't get cheaper just because the AI image was gorgeous

E-commerce images usually get a quick scroll on a phone: small file, short attention span, and buyers lean on the main image plus selling points and reviews to judge. A printed catalog gets handed across a desk by a sales rep, and the client lingers on the same page comparing three, five, sometimes ten SKUs

That's the difference

E-commerce images can get away with a little drama. Catalog images need stable, dependable presentation

・E-commerce can get away with: a brighter background, softer shadows, edges that don't read clearly at thumbnail size. The job is to win the click

・Catalog has to watch out for: on a shared page, mismatched shadow angles, proportion gaps, and color saturation all show up at once. The job shifts to helping a buyer make a comparison

・E-commerce setup: each image stands alone, one product, one frame

・Catalog setup: same-line SKUs sit next to each other. Bottle height, packaging thickness, material sheen all get compared

A case I see all the time: three SKUs in the same bath line. AI tightens one bottle's shoulder line and pumps up the highlight. Solo, it looks refined. Drop it into a catalog and it reads as a different capacity or a different bottle stock

When a buyer sees a spread like that, the first question usually isn't price. It's 'are these three really from the same line?'

For product proportions, set a reference first, then place the items on the layout for comparison. When multiple products share a page, don't trust the visual size the AI gave you. At minimum, run a manual pass using capacity, real dimensions, or the series relationship

I ask designers to lock down three checkpoints first: height, width, and visible footprint

Those three points beat 'looks about right' every time

・Same-line products: use real height or capacity as the proportion anchor. 250ml, 500ml, and 1000ml can't all sit at roughly the same height

・Different categories: decide what the page comparison is actually for, size, appearance, or just lifestyle pairing

・Packaged vs. unpackaged items: keep the box, bottle, and accessories as separate items. Don't let an AI composite push the accessory in front of the hero product

・Hand-held or scene shots: hands, tabletops, cups, plates, and background props all hint at scale. Get those hints wrong and the product scale goes wrong with them

The worst thing a catalog can be is 'proportions look pretty but the sales pitch falls apart.'

Take hardware parts, food gift sets, skincare bundles, clients often use catalogs for first-pass purchasing. Wrong proportions turn the quote, the inventory count, and the on-site conversation into a mess at the same time

When the Mai Strategy Knowledge Academy consulting team helps clients sort catalog assets, we usually recommend doing one 'proportion proof page' first, same-line products laid out together on a single A4 or A3 sheet

That step doesn't take long, but it catches the scale errors AI images are best at hiding

On shadows, check direction, density, and whether the product sits on the ground. On edges, check cut-out fuzz, semi-transparent halos, and tiny holes. On background, check color temperature, brightness, and whether the horizon line is consistent

Add those three together and you usually get the answer to whether AI product photos 'look like they were shot in the same place.'

I open at 100% to read the edges, then zoom back out to read the mood of the full page

Edge problems only show when you zoom in. Shadow and background problems only show when you read the whole page

・Shadow direction: when five products share a page, the light can't come from upper-left on one and lower-right on another, unless the layout is clearly set up as separate scenes

・Shadow density: a clear bottle, a metal tin, and a matte paper box don't cast identical shadows. AI loves to flatten them all into the same soft-focus gray

・Ground contact: the bottom of the product can't float. Bottles, boxes, and small appliances especially need to sit down

・Cut-out edges: on white-background products, check for gray fringe. On dark products, check for light fringe. Felt, brushed surfaces, and clear plastic are the worst offenders for broken edges

・Background consistency: same-page white shouldn't come in three flavors. Once it's on paper, a yellow, blue, or gray drift shows up much faster than it did on screen

A lot of AI retouching pushes product edges too clean

Real products have thickness, bevels, small reflections. An edge that looks laser-cut reads like a sticker

If this catalog run is going through mid-to-high-end commercial printing, a fully custom shop like Minds Printing cares more about edge and background consistency at the proofing stage

Paper stock, varnish, matte film, or spot-UV all magnify the small differences in the image

AI retouching loves to push products brighter, more saturated, and smoother than they really are. In a catalog, the color job is to help people buy the right thing, not to make the image beat the product

Fabric, wood furniture, metal hardware, and food packaging are the four categories where over-retouched material kills trust the fastest

Core term first: color management is the workflow that keeps your screen, file, proof, and printed output as close to each other as possible. The usual work, calibrating the display, assigning an ICC Profile, confirming the CMYK conversion, and proof-comparison, all pushes toward closing the color gap

Before AI product photos hit a catalog, I ask for at least one 'screen-to-paper' mental check

Don't just ask whether it looks good. Ask whether it'll still look like the product once it's printed

・Fabric: texture can't be sanded down to plastic. Weave, nap, and weight still have to read

・Wood grain: pattern can't repeat like a tile. Light and dark layers should sit close to the real piece

・Metal: highlights can't blow out into a flat white sheet. The reflection has to follow the product's shape

・Clear materials: bottles, acrylic, glass have to keep their refraction and thickness. A ring of bright edge isn't enough

・Food packaging: brand color can't drift off the real product just to look punchier. Red, orange, and green are the usual suspects

Print catalogs usually run through a CMYK pipeline, and a lot of the fluorescent, bright blue-green, and high-saturation orange-red that pops on an RGB screen pulls back once it's on paper

If the AI image was already pushed to the limit, the color conversion usually ends one of two ways: it goes dull, or the hue drifts

My method is old-school, but it works: real product, proof, and screen all sit on the same table

Don't mix up the lighting source. Read the main color first, then the shadow, then the material

That gets closer to print reality than dragging sliders on screen for an hour

・Before AI product photos go into a catalog, check believability first. Aesthetics come second

・E-commerce images chase the click. Printed catalogs carry the weight of the buyer's decision

・When multiple products share a page, correct the proportions. The visual size AI gave you is not the truth

・Cut-out edges, shadow direction, and background white are the most common weak points that proofing blows open

・Don't push color past the real product. A catalog sells trust, not a filter

When AI gets pulled into the catalog workflow, the move that pays off isn't handing over all the retouching, it's standardizing the check gates. The design side builds a proportion proof page. Pre-press checks resolution, edges, and CMYK conversion. Sales confirms whether product color and material would mislead the customer. And a SaaS system can turn those fields into required checks before any file uploads

If a small or mid-sized business doesn't have a full workflow yet, start with Minds Printing (MS) three-stage proofing: split every batch of AI product photos into three buckets, ready to drop in, needs retouch correction, needs reshoot or rebuild. Clean assets on the front end are what give the layout, proofing, and printing stages room to talk about quality