ChatGPT IA Guide: Review Dark camisa Tonal Changes

Ask the assistant to compare named changes instead of judging which image looks more luxurious. You are building an acceptance note for an edit, not asking it to certify the shirt's true colour from a screen. A structured review can keep a useful button-area correction while rejecting global lightening. It should distinguish direct observations from interpretations such as shiny, premium or a different fabric.

Prepare the input worksheet

Copy the relevant observations into a small labelled record before prompting. Keep source facts separate from your preferences and requested changes. Replace missing values with unknown rather than inviting the assistant to complete them from general knowledge.

• Confirmed real dark-shirt image
• Three local details ranked by importance
• Original and proposed edit at equal size

Fictional input: the original shows a deep dark torso and a faint button line. A proposed edit makes the buttons clearer but also makes the entire chest look lighter. The sleeve and background were not part of the request.

Copyable master prompt

Create a text brief and review worksheet; do not perform an image edit. Use this requested task as context: Preserve the supplied dark shirt's overall colour and surface appearance. Improve readability around the existing button line using only the local light and details supported by the photograph. Keep the original button count and spacing. Do not brighten the whole torso, add a glossy finish, deepen every fold or create new stitching. Leave the background, pose and other garment regions unchanged. Flag details that cannot be confirmed from this reference. Return separate rows for overall torso tone, button count and spacing, local readability and unintended changes. Recommend accept, revise or ask for evidence for each row. Use only the worksheet supplied below. Keep unknowns visible and do not add product specifications, prices, measurements or policies that are absent from it.

Filled fictional worked example

The following response is an illustrative teaching example. It is not a model test, a TryBuy product measurement or a claim that the assistant verified the physical garment.

This hypothetical output is useful because it does not treat a brighter picture as an automatic improvement. It identifies a conditional success and a separate failure. If the assistant calls the garment black when your supplied listing only says dark, correct the label. The review must retain the source's wording.

Follow-up revision prompts

Use one focused correction at a time. Paste the relevant row or sentence from the response, then identify what must change and what should remain intact.

Rewrite the recommendation as a local correction request: preserve verified button readability while restoring the torso reference. Do not request a full-image brightness increase.

Remove any unsupported material or luxury judgement. Keep only observations about tone, spacing and the requested image region, with uncertainty labelled.

Review the output before using it

• Overall tone unchanged
• Button count and spacing supported
• No invented gloss

Check that accept does not apply to the entire image simply because one feature improved. Each approval must name the region and the evidence supporting it.

Limits and the next real-world step

If the reference does not establish the button construction clearly enough, do not approve a newly detailed button row as fact. Request the relevant real close-up or use another genuine photograph.

The assistant can organise the information you provide; its response is not independent evidence of current stock, exact fit, material performance or store terms. Check the actual listing or physical garment for those decisions. For an image task, written observations can supply the worksheet; this guide does not require or promise an automatic image-editing feature.

Save the final checked worksheet with its source notes, not only the last chat reply. When a fact changes, update that field and ask for the affected part to be revised. Keep the earlier record labelled so that an outdated observation does not quietly return as current evidence.

Use the practical method in preto camisa IA Edits: Reveal Buttons Without Turning the Fabric cinza to collect the real observations first. For related background, consult the relevant TryBuy shirt guide, then return to this worksheet for the specific IA-assisted deliverable.