Hire an AI product photo set reviewer

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• A truth sheet and lock lines, saved and reused for this product
• A planned set with one prompt per frame
• A pass or fail for every picture that comes back, with the smallest fix
• Failed frames redone, or dropped with a note to take a real photo
• A list of faults to add to your prompts next time

How it does the job well

• Start from the user problem and the evidence for it, then the solution.
• Write decisions down with the reason and the options you rejected.
• Keep scope small enough to ship and learn. Name what is out of scope.
• Define how success will be measured before building.

Mistakes it avoids

• Judging a picture from a description or a small preview.
• Redoing one frame again and again. Two rounds, then drop it and take a real photo.
• Adding a word, a price or a claim to a picture because it looks good.
• Calling a set done because the pictures look nice. They must match the real product.
• Running this for a product with no photos yet. Take the photos first.
• Treating opinions or one loud request as proven demand.

What a good result looks like

An example, not your task: You are making product photos with AI and need them to match the real product.

• Truth sheet. What your product really looks like, with the lock lines every prompt repeats and the shots still missing.
• The set. Each passing picture in upload order, with its job and check result.
• Redone. Each frame that was redone, its fault, the fix line and how many rounds it took.
• Dropped. Frames that failed twice, with the fault and the real photo to take.
• Pattern found. Any fault on three or more frames, and the lock line change that fixes it.
• Marketplace result. Pass or reject for the main image, if you sell on a marketplace.

How the work is checked

• Every picture in the set passed the checker, or is listed under dropped.
• Every prompt repeats the lock lines word for word.
• No picture shows a product part, a pack, colour or word that is not on the truth sheet.
• No frame was made that the truth sheet listed as blocked.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

A second opinion

Before it hands back, your AI reviews the draft the way Jakob Nielsen would. This is a way of reviewing the work. It does not mean they wrote or endorse it.

Capabilities

• Get the truth sheet. Writes what your product really looks like from your photos, with the lock lines every prompt repeats. Reuses it next time. Lists the shots it cannot make yet.
• Read the mistakes list. Opens faults from earlier sets and puts the lines that fixed them into every prompt from the start.
• Choose the path. For ads, social or email, it plans a matching shoot. For a product page, it plans about 8 frames that answer shopper questions.
• Hand over the prompts. Gives you each prompt, says how the pictures get made and asks before any credits are used.
• Check every picture. Reads each picture against the truth sheet and the plan. Checks people and hands too, and marketplace rules for the main image, if you say yes to the cost.
• Redo the failures. Adds the exact fix to the prompt and checks the new picture. Two rounds at most, then it drops the frame.
• Fix a pattern at the source. If one fault hits three frames, it fixes the lock line or style sheet first, then redoes them together.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id product-photo-set-maker only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/product-photo-set-maker/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• Clear photos of the exact product, from more than one side
• The product's facts, such as size, variant and what is in the box
• Your brand brain (colours, photo style, what you never show)
• Where the set goes, such as a product page, a marketplace, or ads, social and email
• Customer reviews, to help choose the product page frames (optional)

Helpful context (optional)

• A photo of the product in a hand, for scale
• The listing rules of the marketplace you sell on

Tools and access

• Claude Code: It can keep the photo-set folder, read the pictures you save there and keep the mistakes list.
• ChatGPT or Gemini: Paste a prompt into either one to get the picture. Free plans have tight daily limits.
• Higgsfield: Paste a prompt into its image tool, or connect it so the agent can make the pictures after you say yes.
• Claude: Reads several product photos at full size and quotes printed words line by line with the photo they came from.
• Google Sheets: Keeps one row per variant, so the sheet and its lock lines are easy to find and update.
• Nano Banana: Takes up to 6 product photos plus 3 style references, which suits holding one look across nine frames.
• GPT Image: Keeps the product close to your reference photos, and holds small edits across rounds.
• Any good chat AI: The scoring is simple sums, so any current chat AI can run it. Attach the photos so it can sort them by job.
• Google Sheets: Keeps the score table in rows, so you can re-sort it when a new photo or fact arrives.
• Nano Banana: Built for brand consistency and handles words and layout well, which suits images with short lines on them. Paid plans or credits may apply.
• GPT Image: Lets you point at one word or part on an image and fix it, which suits the one-change remake rule. Paid plans or credits may apply.
• Shopify: Lets you add alt text to each product image in the admin, so the alt lines go straight in.
• Any good chat AI: Any current chat AI that reads images can run the checks. Attach each file at full size, then zoom in yourself too.
• Google Merchant Center: Its Diagnostics page flags product images that break Google's rules, such as promo text or a missing AI label.

Functions and deliverables

• Product photo set reviewer: Plan AI product photos from real references. Check returned images against the product and track fixes or dropped frames.
  • A truth sheet and lock lines, saved and reused for this product
  • A planned set with one prompt per frame
  • A pass or fail for every picture that comes back, with the smallest fix
  • Failed frames redone, or dropped with a note to take a real photo
  • A list of faults to add to your prompts next time
• Product image spec writer: Check your product photos and write a product image spec: a sourced truth sheet, lock lines for every prompt and the next photo to take.
  • A photo check saying which photos to trust and why the rest are out
  • A one-page truth sheet in seven parts, with a source on every line
  • Three to five lock lines that every prompt repeats word for word
  • The shots that can't be made yet, and the one photo to take next
  • Any check the sheet still missed, one line each
• Product photo shoot planner: Plan a product photo shoot where every picture matches. Give it your product photos and colours. You get a style sheet and a prompt for each frame.
  • A six-line style sheet for the whole set
  • A frame list, each with its job, distance, angle and what it shows
  • One prompt per frame, plus a check of the whole set side by side
  • An optional plan for a 6 to 8 second loop clip
  • Any check the set still missed, one line each
• Product page gallery planner: Plan your product page gallery image by image. Give it product facts, reviews and photos. You get an 8-image plan with sourced words and a gaps list.
  • A score table showing why each image made the cut, or didn't
  • One card per image in upload order (question, message, photo, layout, words with sources)
  • A gaps list of every missing fact and photo, and which image waits on each
• Product gallery image planner: Share exact product photos, brand facts and customer feedback. Get a gallery review, image plan and checked files when image tools are available.
  • An audit of your current gallery and the questions it leaves open
  • Customer worries grouped by fear, counted, with real quotes and the image that answers each
  • The image-by-image plan and a gaps list
  • Finished images in upload order, in each shape, with alt text and check results
• Product page image checker: Check your product page images before you upload. Get a pass or fail for each image, one fix per fault and a score out of 100 for the set.
  • PASS or FAIL for each image
  • One line per fault, with the check, severity, where it is, what was planned, what came out and the fix
  • A verdict for the whole set, and a list of any check that couldn't run
  • A score out of 100 for the set, now and after the fixes, with the sum shown
• AI product photo inspector: Share an AI-made product photo. Get eight visual checks, a risk rating and specific prompt fixes, with human review before release.
  • A low, medium or high risk rating
  • A pass or fail result for all eight checks
  • A run, redo or retouch call
  • One to three prompt lines to fix the faults
  Source: dLazy.
• Marketplace photo rule checker: Share product photos and the marketplace. Check image size, colour and layout, then fix supported issues in separate copies.
  • A pass or reject for each photo
  • A plain list of problems
  • Fixed copies where possible
  • A batch summary
  Source: dLazy.
• Product context file writer: Share product facts, audience and positioning. Get a reviewed context file that other marketing skills can reference.
  • A product marketing context file
  • A summary of what the file already holds
  • A change log of updates
  • Gaps you still need to fill
  Source: Corey Haines.
• Jakob Nielsen: Review a website, app or form through Jakob Nielsen’s recorded approach. Find likely usability problems and plan small rounds of realistic user tasks.
  • Your problem worked on the way Jakob Nielsen would: their questions, their method, their push-back
  • The answer or the work, in Jakob Nielsen's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step

Tasks you can use it for

• You are making product photos with AI and need them to match the real product.
• You need a matching set for a product page, a marketplace listing, or ads and social.
• Your last AI set had a wrong label, a warped pack or odd hands, and you do not want that again.

Limits to check before starting

• You have no photos of the product yet: take real photos first, because every check depends on them.
• You only want a plan for a product page with no checks: use the Product page gallery planner.
• You want the agent to draw the pictures itself: it cannot, so make them in ChatGPT, Gemini or Higgsfield and bring them back.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Start a chat with the agent. In Claude Code, add it as a skill or save it in .claude/agents, then ask for a photo set. Without file access, paste the truth sheet and the mistakes list back in at the start of each round.
• Attach clear photos of your real product and say where the set goes.
• Read the plan. For a product page, say yes before it writes the prompts.
• Make the pictures in a tool you pick, such as ChatGPT, Gemini or Higgsfield. The agent cannot draw, and asks before anything uses credits.
• Attach the pictures back so the agent can check each one at full size.
• Redo the frames it sends back. For frames it drops, take a real photo.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/product-photo-set-maker for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.