From Phone Photo to Campaign Visual: The Full Process

In short

• Everything starts from one clear reference photo of the actual product.
• The reference does not need to be well lit — it needs to be sharp, complete and honest about proportions.
• Art direction happens before generation, not after; this is where the quality is decided.
• Typical turnaround is 24 hours for a first set, with one revision round included.

People assume the interesting part is the generation. It is not — the generation is the fastest stage. Here is where the time and the quality actually come from.

Stage 1 — The reference photo

Everything depends on this and it takes you two minutes. We need one sharp photo of the actual product:

• Whole piece visible, nothing cropped off
• Plain surface — a sheet of paper is fine
• Daylight, near a window, no direct flash
• In focus. This is the only non-negotiable.

It does not need to be beautiful. It needs to be truthful: correct proportions, correct colour, every detail legible. A badly lit but sharp phone photo is far more useful to us than a moody, soft one.

Stage 2 — The brief

A short conversation, usually on WhatsApp. What is the product, who buys it, where will these images run, and is there an existing look to match? If you have campaign images you admire — yours or anyone's — this is the moment to send them.

Most quality problems in this industry trace back to a skipped brief. Ten minutes here saves two revision rounds later.

Stage 3 — Art direction

The stage that decides everything, and the one clients never see. Before anything is generated we settle: light direction and quality, palette, surface and staging, camera framing and lens character, and the mood the frame should carry.

This is the same set of decisions a photographer and stylist make on set. Skipping it is what produces the generic, plasticky output that gives AI imagery its bad reputation.

Stage 4 — Generation and selection

Multiple variations are produced against the direction, then culled hard. Most of what is generated is discarded — the same ratio as a conventional shoot, where a photographer might take four hundred frames to deliver twenty.

We are checking for the things that betray a careless result: physically impossible reflections, malformed clasps and chain links, inconsistent shadow direction, proportions drifting from the real product.

Stage 5 — Finishing

Colour grading for consistency across the set, detail retouching, sharpening for the destination medium, and export at the sizes you need — 4:5 for feed, 9:16 for Reels, square for marketplaces, high-resolution for print.

Delivered with full commercial rights. One revision round is included as standard, because the first set is a conversation as much as a deliverable.

Where the time goes

Roughly: ten minutes of your time on the reference and brief, a few hours of art direction and generation, and a couple of hours of finishing. The twenty-four-hour turnaround is mostly queue and review, not compute.

What a good reference photo looks like

Since this determines everything, it is worth being precise. A strong reference photo has:

• Even, indirect daylight. A window with a sheer curtain, or open shade outdoors. Never direct sun, never on-camera flash.
• A plain surface. White or mid-grey paper. Not patterned fabric, not a wooden table with strong grain.
• The whole product in frame with a small margin, nothing cropped, nothing overlapping.
• Sharp focus on the product's defining detail — the setting, the clasp, the weave.
• No filters, no beautification, no HDR. Phone camera defaults often over-sharpen and over-saturate; turn those off if you can.

Three additional frames improve the outcome noticeably: one at a 45-degree angle, one from directly above, and one macro on the key detail.

What happens during art direction, concretely

For a single necklace, the decisions settled before generation typically include: key light direction and softness; whether there is a visible specular highlight and where it falls; the surface material and its texture scale; the palette in three specified colours; camera height and apparent focal length; depth of field; whether there is a human element and how much of it is in frame; and the emotional register — quiet and archival, or warm and celebratory.

Written down, that is roughly a paragraph per image family. It is the same paragraph a photographer holds in their head walking onto a set, and skipping it is exactly why casual AI output looks casual.

The review checklist before anything is sent

1. Does the product match the reference in proportion, colour and detail?
2. Hands: correct count, plausible joints, believable contact with the object?
3. Closures and hardware: mechanically possible?
4. Shadows: consistent direction and softness across the frame?
5. Reflections: could the reflected content actually be where the scene implies?
6. Text and marks: composited from the reference, never generated?
7. Does it hold up at 100% zoom, not just as a thumbnail?

Revisions

One round is included and most clients use it. The most common requests are practical rather than creative — warmer grade, more negative space at the top for a text overlay, a tighter crop for a marketplace slot. Sending your intended layout with the brief avoids most of them entirely.

Frequently asked

What kind of photo do I need to send?

A sharp, well-focused photo of the whole product against a plain surface, taken in daylight near a window. Phone cameras are entirely adequate. What matters is focus, complete visibility of the piece and accurate proportions — not lighting quality.

How long does the whole process take?

A first set typically lands within 24 hours of the brief being agreed. Revisions add a few hours each. Larger catalogues are batched and scheduled, but the per-image time does not change much.

Can you match a specific campaign look I've seen?

Yes. Send the reference image alongside your product photo. Matching an existing look is usually easier than working from a written description, because it removes ambiguity about light, palette and mood.