Orange Pyramid
Purple Sphere
Blue Cylinder
Turquoise Star
Lime Green Object
Yellow Cube
Yellow Cube

Gil & Roy Props

Fixing a fragmented workflow, one bottleneck at a time.

Gil & Roy Props is a 100,000-sq-ft commercial prop house in Los Angeles. Serving major film and TV productions, the store has worked with productions ranging from student films to Men in Black.


While working the front desk, I noticed that sales reps, dock workers, and accountants had to juggle an antiquated Shopify system with paper invoices, physical tags, and spreadsheets.


Rather than attempting a major overhaul, I engineered a suite of tools to resolve our biggest day-to-day hurdles.

ROLE

Workflow Analysis · Product Strategy · UX/UI · Design Engineering · Front-End Development

TEAM

4 Sales Reps · 5 Dock Workers · Drivers · Accountant


TIMELINE

Iterative Internal Deployments

TOOLS

Next.js · React · TypeScript · Tailwind CSS · PostgreSQL · SQLite · Supabase · Prisma · Google Gemini API · Anthropic Claude API · OpenAI API · AWS S3 · Vercel · Fly.io · Shopify · Chrome Extensions

MY CHALLENGE:

How do you eliminate operational bottlenecks when you can't rebuild a legacy system from scratch?

I Problem

When renting a prop is not just renting a prop.

At Gil & Roy, with any given order, props could move across many channels, including our Shopify website, phone calls, physical tags, or emails.

During an order's lifespan, sales reps (including myself) would pause customer interactions to hunt down inventory while juggling various threads of communication.

At the same time, dock workers were responsible for grabbing orders without a centralized view of what was being pulled, staged, or loaded for pickup.

Since our core Shopify app was maintained by a third-party developer, the full system was out of our hands.

Paying close attention to each bottleneck, I decided to work around the core system instead of through it.

With my background in frontend development, I identified key friction points on the floor and engineered four tools to fix our biggest bottlenecks.

II Process: Four Targeted Solutions

Process: Four Targeted Solutions

01. THE Prop Ticker

On the loading dock, workers relied on paper tickets and memory, constantly walking across a 100,000-sq-ft warehouse to confirm every order's status.

Each day, I watched the team:

  • Repeat similar tasks

  • Search endlessly for information

  • Ask one another identical questions

  • Walk back and forth to find the same items

  • Rely too much on personal memory

It became clear we needed a single, visual framework of warehouse orders.

Luckily, an astute coworker brought my attention to an unused but serviceable LED TV on a shelf overlooking the docks.

I would leverage this as a visual framework — my first big breakthrough.

THE GIL & ROY PROP TICKER

💻 Next.js · Supabase · Tailwind CSS · Google Calendar API · Vercel

1

3

2

THE GIL & ROY PROP TICKER

REAL TIME ORDER TRACKING

1

DOCK LOCATION

2

ORDER STATUS

3

3

2

1

1

3

2

SYNCED WITH GOOGLE CAL API

1

DAY OF AND NEXT DAY UPDATES

2

DELIVERY STATUS

3

02. THE AI PRODUCT BUILDER

Once the docks had a clearer picture of the day's workflow, I was emboldened to tackle another key issue — our incomplete catalog.

Creating new products was drastically slowing down order processing!

While the prop house's digital catalog was supposed to contain every item in the warehouse, we always encountered orders with new items.

Unfortunately, creating multiple items in Shopify is a tedious process.

Few of the sales reps understood how to navigate the sales catalog, let alone create photos and assign important SKU numbers to new items.

THE GIL & ROY AI PRODUCT BUILDER

💻 Next.js · PostgreSQL + Prisma · Google Gemini API · OpenAI API · Shopify CSV Import · Cloudflare R2

IN ADDITION TO THE TICKER AND PRODUCT BUILDER I MADE TWO MORE TOOLS FOR OUR WAREHOUSE

03. THE ACCOUNT TRACKER

Every day, I witnessed the team's accountant fill out a curated spreadsheet with customer account information.*

*For some background, any buyer needed to provide a few pieces of information to get an account started.

The most important piece of information for a new customer was a Certificate of Insurance (COI). This document protected both the house and customer from being liable for loss or damages when renting props.

From this single PDF, the accountant could extract:

  • Billing information

  • Contact information

  • Whether the COI was current

As good as the accountant's spreadsheet was at tracking information, it had room for improvement.

I built an Account Tracker with Next.js, Prisma, and SQLite that uses the Anthropic Claude API to extract COI details from PDFs automatically.

04. THE EOD REPORTING WIDGET

Finally, I built a tool to address a much smaller (but extremely repetitive) task.

At the end of every day, the accountant needed to compile photo IDs associated with orders for end-of-day reporting.

This information existed in our third-party Shopify app, but it was built in such a way that the user had to manually view each order and click a link to generate an invoice PDF.

To make things easier, I developed a Chrome extension to streamline the process of gathering the photo information for the accountant's daily reporting.

The widget was a simple Chrome Extension built with vanilla JavaScript.

IV Final Thoughts

Big systems don't always need big solutions.

Rather than forcing an extensive platform migration, I built a modular suite of tools to help our staff move more quickly and efficiently.


At the end of the process I had:


  • Reduced multiple channels of communication to a single, shared status screen on the loading dock.

  • Streamlined inventory creation using multimodal vision AI.

  • Automated COI verification and account generation with AI document extraction.

  • Eliminated end-of-day report bottlenecks with batch invoice processing.

Key Takeaways

  1. Think critically about the workspace.


  1. Respect existing mental models.


  1. Don't force a platform rebuild if simpler solutions are more attainable.


  1. Keep AI practical and unobtrusive; don't lean on it for every answer.

METRICS

  • 1,000+ props cataloged with the AI Product Builder, in batches instead of one Shopify listing at a time

  • 20+ orders a day shown on the Prop Ticker to a 12-person team, with fewer missed pulls

  • 75% faster end-of-day reporting, from about an hour to 15 minutes

  • Seconds instead of 10–15 minutes to verify insurance across ~800 client accounts

  • Supports a warehouse doing ~$5M in annual film and TV sales

THE GIL & ROY PRODUCT HELPER

REMOVE BACKGROUND BEFORE UPLOADING

1

1

1

3

2

EDIT NEW PROP META DATA

1

AI GENERATES CONTENT

2

AVAILABLE SKU FROM INDEX

3

Orange Pyramid
Purple Sphere
Blue Cylinder
Turquoise Star
Lime Green Object
Yellow Cube
Yellow Cube

Gil & Roy Props

Fixing a fragmented workflow, one bottleneck at a time.

Gil & Roy Props is a 100,000-sq-ft commercial prop house in Los Angeles. Serving major film and TV productions, the store has worked with productions ranging from student films to Men in Black.


While working the front desk, I noticed that sales reps, dock workers, and accountants had to juggle an antiquated Shopify system with paper invoices, physical tags, and spreadsheets.


Rather than attempting a major overhaul, I engineered a suite of tools to resolve our biggest day-to-day hurdles.

ROLE

Workflow Analysis · Product Strategy · UX/UI · Design Engineering · Front-End Development

TEAM

4 Sales Reps · 5 Dock Workers · Drivers · Accountant


TIMELINE

Iterative Internal Deployments

TOOLS

Next.js · React · TypeScript · Tailwind CSS · PostgreSQL · SQLite · Supabase · Prisma · Google Gemini API · Anthropic Claude API · OpenAI API · AWS S3 · Vercel · Fly.io · Shopify · Chrome Extensions

MY CHALLENGE:

How do you eliminate operational bottlenecks when you can't rebuild a legacy system from scratch?

I Problem

When renting a prop is not just renting a prop.

At Gil & Roy, with any given order, props could move across many channels, including our Shopify website, phone calls, physical tags, or emails.

During an order's lifespan, sales reps (including myself) would pause customer interactions to hunt down inventory while juggling various threads of communication.

At the same time, dock workers were responsible for grabbing orders without a centralized view of what was being pulled, staged, or loaded for pickup.

Since our core Shopify app was maintained by a third-party developer, the full system was out of our hands.

Paying close attention to each bottleneck, I decided to work around the core system instead of through it.

With my background in frontend development, I identified key friction points on the floor and engineered four tools to fix our biggest bottlenecks.

II

Process: Four Targeted Solutions

01. THE Prop Ticker

On the loading dock, workers relied on paper tickets and memory, constantly walking across a 100,000-sq-ft warehouse to confirm every order's status.

Each day, I watched the team:

  • Repeat similar tasks

  • Search endlessly for information

  • Ask one another identical questions

  • Walk back and forth to find the same items

  • Rely too much on personal memory

It became clear we needed a single, visual framework of warehouse orders.

Luckily, an astute coworker brought my attention to an unused but serviceable LED TV on a shelf overlooking the docks.

I would leverage this as a visual framework — my first big breakthrough.

THE GIL & ROY PROP TICKER

💻 Next.js · Supabase · Tailwind CSS · Google Calendar API · Vercel

1

3

2

THE GIL & ROY PROP TICKER

REAL TIME ORDER TRACKING

1

DOCK LOCATION

2

ORDER STATUS

3

3

2

1

1

3

2

SYNCED WITH GOOGLE CAL API

1

DAY OF AND NEXT DAY UPDATES

2

DELIVERY STATUS

3

02. THE AI PRODUCT BUILDER

Once the docks had a clearer picture of the day's workflow, I was emboldened to tackle another key issue — our incomplete catalog.

Creating new products was drastically slowing down order processing!

While the prop house's digital catalog was supposed to contain every item in the warehouse, we always encountered orders with new items.

Unfortunately, creating multiple items in Shopify is a tedious process.

Few of the sales reps understood how to navigate the sales catalog, let alone create photos and assign important SKU numbers to new items.

THE GIL & ROY AI PRODUCT BUILDER

💻 Next.js · PostgreSQL + Prisma · Google Gemini API · OpenAI API · Shopify CSV Import · Cloudflare R2

IN ADDITION TO THE TICKER AND PRODUCT BUILDER I MADE TWO MORE TOOLS FOR OUR WAREHOUSE

03. THE ACCOUNT TRACKER

Every day, I witnessed the team's accountant fill out a curated spreadsheet with customer account information.*

*For some background, any buyer needed to provide a few pieces of information to get an account started.

The most important piece of information for a new customer was a Certificate of Insurance (COI). This document protected both the house and customer from being liable for loss or damages when renting props.

From this single PDF, the accountant could extract:

  • Billing information

  • Contact information

  • Whether the COI was current

As good as the accountant's spreadsheet was at tracking information, it had room for improvement.

I built an Account Tracker with Next.js, Prisma, and SQLite that uses the Anthropic Claude API to extract COI details from PDFs automatically.

04. THE EOD REPORTING WIDGET

Finally, I built a tool to address a much smaller (but extremely repetitive) task.

At the end of every day, the accountant needed to compile photo IDs associated with orders for end-of-day reporting.

This information existed in our third-party Shopify app, but it was built in such a way that the user had to manually view each order and click a link to generate an invoice PDF.

To make things easier, I developed a Chrome extension to streamline the process of gathering the photo information for the accountant's daily reporting.

The widget was a simple Chrome Extension built with vanilla JavaScript.

IV Final Thoughts

Big systems don't always need big solutions.

Rather than forcing an extensive platform migration, I built a modular suite of tools to help our staff move more quickly and efficiently.


At the end of the process I had:


  • Reduced multiple channels of communication to a single, shared status screen on the loading dock.

  • Streamlined inventory creation using multimodal vision AI.

  • Automated COI verification and account generation with AI document extraction.

  • Eliminated end-of-day report bottlenecks with batch invoice processing.

Key Takeaways

  1. Think critically about the workspace.


  1. Respect existing mental models.


  1. Don't force a platform rebuild if simpler solutions are more attainable.


  1. Keep AI practical and unobtrusive; don't lean on it for every answer.

METRICS

  • 1,000+ props cataloged with the AI Product Builder, in batches instead of one Shopify listing at a time

  • 20+ orders a day shown on the Prop Ticker to a 12-person team, with fewer missed pulls

  • 75% faster end-of-day reporting, from about an hour to 15 minutes

  • Seconds instead of 10–15 minutes to verify insurance across ~800 client accounts

  • Supports a warehouse doing ~$5M in annual film and TV sales

THE GIL & ROY AI PRODUCT BUILDER

REMOVE BACKGROUND BEFORE UPLOADING

1

1

1

3

2

EDIT NEW PROP META DATA

1

AI GENERATES CONTENT

2

AVAILABLE SKU FROM INDEX

3

Orange Pyramid
Purple Sphere
Blue Cylinder
Turquoise Star
Lime Green Object
Yellow Cube
Yellow Cube

Gil & Roy Props

Fixing a fragmented workflow, one bottleneck at a time.

Gil & Roy Props is a 100,000-sq-ft commercial prop house in Los Angeles. Serving major film and TV productions, the store has worked with productions ranging from student films to Men in Black.


While working the front desk, I noticed that sales reps, dock workers, and accountants had to juggle an antiquated Shopify system with paper invoices, physical tags, and spreadsheets.


Rather than attempting a major overhaul, I engineered a suite of tools to resolve our biggest day-to-day hurdles.

ROLE

Workflow Analysis · Product Strategy · UX/UI · Design Engineering · Front-End Development

TEAM

4 Sales Reps · 5 Dock Workers · Drivers · Accountant


TIMELINE

Iterative Internal Deployments

TOOLS

Next.js · React · TypeScript · Tailwind CSS · PostgreSQL · SQLite · Supabase · Prisma · Google Gemini API · Anthropic Claude API · OpenAI API · AWS S3 · Vercel · Fly.io · Shopify · Chrome Extensions

MY CHALLENGE:

How do you eliminate operational bottlenecks when you can't rebuild a legacy system from scratch?

I Problem

When renting a prop is not just renting a prop.

At Gil & Roy, with any given order, props could move across many channels, including our Shopify website, phone calls, physical tags, or emails.

During an order's lifespan, sales reps (including myself) would pause customer interactions to hunt down inventory while juggling various threads of communication.

At the same time, dock workers were responsible for grabbing orders without a centralized view of what was being pulled, staged, or loaded for pickup.

Since our core Shopify app was maintained by a third-party developer, the full system was out of our hands.

Paying close attention to each bottleneck, I decided to work around the core system instead of through it.

With my background in frontend development, I identified key friction points on the floor and engineered four tools to fix our biggest bottlenecks.

II

Process: Four Targeted Solutions

01. THE Prop Ticker

On the loading dock, workers relied on paper tickets and memory, constantly walking across a 100,000-sq-ft warehouse to confirm every order's status.

Each day, I watched the team:

  • Repeat similar tasks

  • Search endlessly for information

  • Ask one another identical questions

  • Walk back and forth to find the same items

  • Rely too much on personal memory

It became clear we needed a single, visual framework of warehouse orders.

Luckily, an astute coworker brought my attention to an unused but serviceable LED TV on a shelf overlooking the docks.

I would leverage this as a visual framework — my first big breakthrough.

THE GIL & ROY PROP TICKER

💻 Next.js · Supabase · Tailwind CSS · Google Calendar API · Vercel

1

3

2

THE GIL & ROY PROP TICKER

REAL TIME ORDER TRACKING

1

DOCK LOCATION

2

ORDER STATUS

3

3

2

1

1

3

2

SYNCED WITH GOOGLE CAL API

1

DAY OF AND NEXT DAY UPDATES

2

DELIVERY STATUS

3

02. THE AI PRODUCT BUILDER

Once the docks had a clearer picture of the day's workflow, I was emboldened to tackle another key issue — our incomplete catalog.

Creating new products was drastically slowing down order processing!

While the prop house's digital catalog was supposed to contain every item in the warehouse, we always encountered orders with new items.

Unfortunately, creating multiple items in Shopify is a tedious process.

Few of the sales reps understood how to navigate the sales catalog, let alone create photos and assign important SKU numbers to new items.

THE GIL & ROY AI PRODUCT BUILDER

💻 Next.js · PostgreSQL + Prisma · Google Gemini API · OpenAI API · Shopify CSV Import · Cloudflare R2

IN ADDITION TO THE TICKER AND PRODUCT BUILDER I MADE TWO MORE TOOLS FOR OUR WAREHOUSE

03. THE ACCOUNT TRACKER

Every day, I witnessed the team's accountant fill out a curated spreadsheet with customer account information.*

*For some background, any buyer needed to provide a few pieces of information to get an account started.

The most important piece of information for a new customer was a Certificate of Insurance (COI). This document protected both the house and customer from being liable for loss or damages when renting props.

From this single PDF, the accountant could extract:

  • Billing information

  • Contact information

  • Whether the COI was current

As good as the accountant's spreadsheet was at tracking information, it had room for improvement.

I built an Account Tracker with Next.js, Prisma, and SQLite that uses the Anthropic Claude API to extract COI details from PDFs automatically.

04. THE EOD REPORTING WIDGET

Finally, I built a tool to address a much smaller (but extremely repetitive) task.

At the end of every day, the accountant needed to compile photo IDs associated with orders for end-of-day reporting.

This information existed in our third-party Shopify app, but it was built in such a way that the user had to manually view each order and click a link to generate an invoice PDF.

To make things easier, I developed a Chrome extension to streamline the process of gathering the photo information for the accountant's daily reporting.

The widget was a simple Chrome Extension built with vanilla JavaScript.

IV Final Thoughts

Big systems don't always need big solutions.

Rather than forcing an extensive platform migration, I built a modular suite of tools to help our staff move more quickly and efficiently.


At the end of the process I had:


  • Reduced multiple channels of communication to a single, shared status screen on the loading dock.

  • Streamlined inventory creation using multimodal vision AI.

  • Automated COI verification and account generation with AI document extraction.

  • Eliminated end-of-day report bottlenecks with batch invoice processing.

Key Takeaways

  1. Think critically about the workspace.


  1. Respect existing mental models.


  1. Don't force a platform rebuild if simpler solutions are more attainable.


  1. Keep AI practical and unobtrusive; don't lean on it for every answer.

METRICS

  • 1,000+ props cataloged with the AI Product Builder, in batches instead of one Shopify listing at a time

  • 20+ orders a day shown on the Prop Ticker to a 12-person team, with fewer missed pulls

  • 75% faster end-of-day reporting, from about an hour to 15 minutes

  • Seconds instead of 10–15 minutes to verify insurance across ~800 client accounts

  • Supports a warehouse doing ~$5M in annual film and TV sales

THE GIL & ROY AI PRODUCT BUILDER

REMOVE BACKGROUND BEFORE UPLOADING

1

1

1

3

2

EDIT NEW PROP META DATA

1

AI GENERATES CONTENT

2

AVAILABLE SKU FROM INDEX

3

© 2026 BEN NICHOLSON