Duplicating a Google Drive folder: we coded the add-on with Claude

Fifteen years of using Google Drive. Then one morning, while trying to duplicate an onboarding folder for a new client, we ran into an absurdity: the tool simply can't do it. Here is the story of a Workspace add-on coded in 30 minutes with Claude, even though we had never written a single line of Apps Script.

Duplicating a Google Drive folder with its contents is not a native feature of the tool, despite its 3 billion monthly active users. To work around this gap, we built an in-house Google Workspace add-on in 30 minutes with Claude, with no prior experience in Google Apps Script. The add-on is now deployed internally at Fragments Studio and lets us copy an entire folder tree in one click.

This article walks through the build step by step and what it reveals about a broader reality: in 2026, AI unlocks internal projects that no company would have prioritized otherwise.

Why can't you duplicate a folder in Google Drive?

Google Drive lets you copy a file with a right-click, but offers no option to duplicate a folder and its contents. It is a documented, long-standing limitation, regularly raised by the community for more than ten years. As early as 2020, The Next Web ran the headline "Why the hell can't you copy folders in Google Drive?", and the question is still open.

The concrete context of the need at Fragments Studio

For every new client, we set up a structured Drive space: folders for scoping, design, dev, meeting notes, deliverables and invoices. It is an onboarding standard that ensures every team finds the same folder tree from one project to the next. The natural reflex: create a template folder, then duplicate it for each new account.

Except Google Drive can't do it. The only manual hack is to download the folder as a .zip, unzip it locally, then re-upload it under a new name. Seven to ten minutes of fiddling, loss of all shared permissions, and the risk of missing a file, for an operation that should take two seconds.

Why this friction deserves attention

An invisible operation repeated twenty times a year adds up to two to three hours lost per year per user on a task with zero value. Multiply that by a team, by HR onboardings, by client project templates, and the hidden cost becomes significant.

Above all, this is exactly the kind of friction no IT department would ever approve as a priority project. Too small for the roadmap, too chronic to forget. That is exactly where AI changes the game.

How did we code the Google Workspace add-on with Claude in 30 minutes?

A Google Workspace add-on is an extension that adds custom features to Google apps (Drive, Docs, Sheets, Gmail). It is written in Google Apps Script, a proprietary language derived from JavaScript, and can be published on the Google Workspace Marketplace, which hosts more than 5,000 apps with more than 5 billion cumulative installs.

Here is exactly how the build went, from the initial roadblock to the deployed add-on.

Step 1: Scope the need and generate the code (Claude chat, ~15 min)

First prompt to Claude: "I want to create a Google Workspace add-on that adds a 'Duplicate this folder' button in Google Drive. What is the fastest stack?"

A clear answer: Apps Script + the native Drive API, deployable straight from the online editor. No toolchain to install, no compilation. Claude then generates the complete code: recursion over the folder tree, copying each file via DriveApp.getFileById().makeCopy(), recreating the folder structure, and error handling for protected files.

Two debugging iterations to handle the case of native Google Docs (which are not copied like binary files), and the code was running.

Step 2: Configure OAuth and deploy (Claude in Chrome, ~15 min)

This is where it gets interesting. Configuring OAuth in the Google Cloud Platform console is notoriously painful: create a project, enable the right APIs, configure the consent screen, manage scopes, submit the add-on to the private marketplace. Half a day of wandering for a beginner.

Instead, we delegated the navigation to Claude in Chrome, Anthropic's browser agent. A single instruction: "Configure this add-on for internal deployment on our Workspace, with the minimum scopes needed to copy folders." The agent opened the GCP console, created the project, enabled the Drive API, configured the consent screen, defined the OAuth scopes, then submitted the add-on to the Fragments private marketplace. Fifteen minutes, without a single configuration error.

The build results in numbers

The add-on is now used internally by the whole team. Publishing it on the public marketplace is being considered for other organizations that face the same friction.

What this build says about internal AI for companies

Beyond the one-off gain, this project illustrates a deep shift. 65% of companies had adopted a citizen development model in 2025, and those that automate these processes see an average 40% reduction in development costs according to Quandary Consulting. Apps built in-house are deployed 5 to 10 times faster than through traditional cycles.

But it is not so much speed that changes the game, it is the category of projects that suddenly becomes feasible. Before AI, a Workspace add-on to fix an internal friction was a project too small to tie up a senior developer, too specific to buy an off-the-shelf solution. So nothing got done.

At Fragments Studio, we see that the best AI use cases in companies almost never come from management. They emerge from the frictions teams experience every day, which they did not even dare frame as projects.

Three conditions to replicate this kind of gain internally

The story above is not a stroke of luck. It rests on three specific conditions that must come together for AI to truly deliver.

1. A precise, quantifiable operational problem. "Duplicate a folder in one click" is a measurable deliverable. "Improve productivity" is not. Good internal AI use cases are always framed as a concrete, repeated task with an identifiable hidden cost.
2. Initial scoping with a human who knows the business. AI cannot guess that you need to preserve Drive permissions during the copy. The initial prompt must contain the real constraints, otherwise the generated code will be elegant but unusable.
3. A human validation step before deployment. An add-on that touches company files must be reviewed by someone who understands the implications. AI speeds up the build, it does not remove responsibility.

These three conditions are precisely what we equip teams with in our AI support for companies. The goal is not to replace your teams with models, but to give them the methods to identify, scope and deliver these small internal projects whose value adds up. For more complex needs, such as business integrations and custom applications, our custom development team takes over with the same philosophy: AI as an amplifier, not a substitute.

Frequently asked questions

How do you duplicate a Google Drive folder with its contents? Google Drive does not offer a native function to duplicate a complete folder. There are three options: copy the files one by one manually, download the folder as a .zip and then re-upload it, or install a dedicated Workspace add-on that performs the recursive copy via Google Apps Script.

What is a Google Workspace add-on? A Google Workspace add-on is an extension that adds custom features to Google apps (Drive, Docs, Sheets, Gmail, Calendar). It is written in Google Apps Script and can be deployed internally on an organization's private marketplace, or published publicly on the Google Workspace Marketplace.

How long does it take to develop a Workspace add-on with AI? For a simple add-on (one function, one button, recursive logic), 30 minutes to 2 hours is enough when you use an AI assistant to generate the Apps Script code and a browser agent to automate OAuth configuration and deployment. Without AI and without prior experience, expect one to two days of work instead.

Is the add-on publicly available on the Google Workspace Marketplace? For now, the add-on is deployed only on the Fragments Studio private marketplace. Publishing it on the public marketplace is being considered so that other organizations can benefit from the feature.

Conclusion

The story of this add-on is not about its code, which remains trivial. It is about what it reveals: the barrier to fixing internal frictions yourself has just been shattered. Thirty minutes to deliver what would have taken a senior developer a day, or, more likely, would never have been prioritized.

The real opportunity in 2026 is not building spectacular AI products. It is methodically eliminating the small, invisible frictions that have been costing all your teams, every week, for years. The tools are here. All that remains is to look at them differently.

Want to identify your own internal AI use cases?

At Fragments Studio, we support companies that want to turn their operational frictions into concrete, measurable projects.

• Discover our AI support for companies to map your use cases and deliver your first internal projects.
• If your need calls for broader development, our custom development team integrates AI into complete business solutions.
• Get an initial estimate with our project estimator.
• Or let's talk directly to scope your first internal AI project.