Save moneyBusiness size: Small team

A Singapore nasi padang chain owner says ChatGPT makes his food marketing images for about $25 a month instead of a $2,000 photo shoot

Eugene Tan · Singapore · ChatGPT

Go to the skill

What the agent did

Eugene Tan, owner of Serangoon BBQ & Curry, photographs the parts of a new dish on his phone and feeds the pictures into ChatGPT along with a reference image. The Straits Times says the marketing material is ready in a few minutes.

Result, as claimed

The Straits Times reports that two years ago he would have budgeted around $2,000 for a full day of photography and several weeks of work. He switched to AI photos in 2025 and now makes the material for around $25 a month. He says a new product can be launched and promoted in as few as five days.

This is the result as claimed in the public source linked below. It has not been checked and is not a typical outcome.

Proof

Read the story in The Straits Times

Try it yourself

Turn phone photos of a dish's parts into one marketing image that stays true to what the customer will get.

Difficulty: Beginner

Prompt

I am attaching phone photos of each part of my new dish [dish name] and one reference image of the style I like [attach or describe]. Make one marketing image for [channel]. Rules: show only the items, portions and colours in my photos. Do not add ingredients, garnish or extra portions. Keep textures natural. Before you generate, list what you will show and anything you are unsure about so I can correct you. After you generate, list anything in the image that differs from my photos.

Prompt: starter by Solopreneur Code

Tools

  • ChatGPT Eugene Tan uses ChatGPT with photos of the dish parts and a reference image

Steps

  1. Photograph each part of the dish separately in good light.
  2. Pick one reference image for the style.
  3. Run the prompt and read the list before the image.
  4. Put the image next to a real plate and compare.
  5. Regenerate until a real plate would look the same, or use your own photo.

Watch out: Two of his customers complained that the image did not match the dish, so check every image against a real plate before it goes live.

Starter version by Solopreneur Code

Skill: Marketing

Add this skill

MarketingDifficulty: Beginner

Honest dish image from component photos

Turns phone photos of a dish's parts into one marketing image that shows only what the customer will get, with a check list of differences.

Works with
ChatGPT, Gemini, Claude Needs an AI chat that can read and make images. Check every image against a real plate.
Tools in the story
ChatGPT
Country
Singapore

How to use this: Copy the skill, paste it into Claude, ChatGPT or Gemini, and ask it to save it as a skill. If it cannot, it will still follow the steps.

Show skill

If the Copy button does not work, select this text and copy it.

Sign up for the newsletter

Get the next story from Solopreneur Code in your inbox.

Form not loading? Subscribe on Substack

Back to all stories

More like this

Browse all stories
Save moneyBusiness size: Small teamIncludes prompt

An 8-person AI company runs written AI playbooks in the background, and its CEO says it would otherwise need four to five more hires

Rachel Woods · Country unknown · Claude Code, Claude, ChatGPT

AI agents run AMP's department playbooks in the background through Claude Code. One reviews sales calls against company standards. Another answers sales team questions from the team's guide.

Result, as claimed

Woods says AMP has four full-time employees and four part-time contractors and would need at least four to five more people without its AI playbooks, including a sales and marketing manager, customer support specialist, executive assistant and operations coordinator. In September it spent $8,510 on Claude Code usage to run the playbooks, on top of team Claude and ChatGPT subscriptions.

Save moneyBusiness size: Small teamIncludes prompt

A 14-truck Tennessee fleet owner built his own trucking back office with Claude Code and ChatGPT for $55 a month and says it saves 20 to 25 hours a week

Wes Oberman · United States · Claude Code, ChatGPT, Claude Sonnet 4.6, Claude Opus 4.8, GitHub, Supabase, Vercel, Resend

Claude Code built the system from prompts ChatGPT drafted. In daily use, Claude Sonnet 4.6, backed up by Claude Opus 4.8, reads a broker rate confirmation PDF and fills the load fields in 15 to 20 seconds, including deduction terms. The system emails load details to the owner operator, tracks status to delivery, and combines invoice, rate confirmation and proof of delivery into one PDF for the broker.

Result, as claimed

Oberman says the system saves 20 to 25 hours a week and costs $55 a month in tool subscriptions, against a cheapest outside quote of about $1,500 a month. A separate AI settlement tool he built cut his wife's Thursday settlements from a 12-hour day to about an hour.

Save moneyBusiness size: SoloIncludes prompt

A Singaporean with no coding skills used Codex to build a tool that picks the best credit card for miles at 3,300 merchants

Hosan Swee (KiasuMiles) · Singapore · Codex, ChatGPT

Hosan Swee and his wife share about a dozen credit cards and kept losing miles by picking the wrong one. He used Codex and ChatGPT to build KiasuMiles. Codex scanned bank and merchant websites to build a structured database of card rules, including minimum spend, miles per dollar, eligible categories and monthly caps. People connect KiasuMiles to ChatGPT, Claude or Codex and ask which card to use at a given shop. He built it feature by feature with specific prompts and tested the logic on a few cards first.

Result, as claimed

The Straits Times reports the tool tracks around 50 credit cards and over 3,300 merchants in Singapore. Compiling the card rules was done in minutes with Codex, compared with the hours it would take by hand. No miles or dollar savings figure is given.