Skip to content
Salah Eddine El-Azhary profile

Salah Eddine El-Azhary

Software Engineer
  • Email salah9azhary9@gmail.com
  • Location Marrakech, MA
About Resume Portfolio Contact Gallery
Case study chromewebstore.google.com
Visit project ×
CASE STUDY · 2025–2026

Building a Pinterest AI Content Filter: From a User Problem to an Organic Product

How I identified the growing problem of AI-generated content on Pinterest, built a browser extension to give users more control over what they see, and developed an organic growth strategy around the problem itself.

Role: Software Engineer & Builder Platform: Chrome & Firefox Started: 2025
01 · CONTEXT

The problem started with a change in visual content

Around 2025, AI-generated images became increasingly present across visual-content platforms. Pinterest, in particular, is built around discovering, saving, and exploring visual ideas, which made the growing presence of AI-generated content noticeable to users.

The problem was not simply that AI images existed. The bigger issue was the lack of a simple way for a user to distinguish between AI-generated images and content created or photographed by people.

THE USER PROBLEM

“I want to browse Pinterest normally, but I don't necessarily want AI-generated images mixed into my visual discovery experience.”

02 · OPPORTUNITY

What if the user could decide what to see?

Instead of building another standalone image platform, I looked at the problem from a different angle:

What if users could continue using Pinterest exactly as they normally do, while having an additional layer of control over AI-generated content?

This led to the idea of building a browser extension capable of helping users filter AI-generated visual content directly while they browse Pinterest.

  • Keep the existing Pinterest experience.
  • Add an additional filtering layer.
  • Give users more control over the content they encounter.
  • Avoid requiring users to move to another platform.
03 · SOLUTION

Building the Pinterest AI Content Filter

I started building the browser extension as the first person responsible for bringing the concept from an idea into a usable product.

The extension was designed for Google Chrome and Mozilla Firefox, allowing users to continue browsing Pinterest while the extension adds the filtering experience on top of the existing website.

01 Pinterest User browses normally
→
02 Browser Extension Processes visible content
→
03 Filtering Layer Helps separate AI content
→
04 User Experience More controlled discovery
04 · PRODUCT THINKING

The interesting part wasn't only building the extension

Building the technical solution was only one part of the project. I also started thinking about a second problem: How do you make people discover a browser extension that solves a very specific problem?

Rather than trying to market the product around a broad concept such as “AI image detection,” I focused on the exact problem users were already searching for.

GROWTH STRATEGY

Own the problem, not just the product

The strategy was to build an organic presence around the search intent behind the problem, particularly the keyword:

Pinterest AI Content Filter

The goal was to make the extension discoverable when users were already looking for a solution to this specific problem.

05 · ORGANIC GROWTH

Turning a product feature into an SEO opportunity

The product itself created a natural SEO opportunity because the problem was specific enough to have clear search intent.

Instead of creating generic content about artificial intelligence, I focused the organic strategy around the relationship between Pinterest, AI-generated images, and content filtering.

User experiences a problem → User searches for a solution → Finds relevant content → Discovers the extension

This created a direct connection between search intent → problem → solution → product.

06 · RESULTS

What happened after launching?

The project gradually started generating organic usage rather than depending entirely on paid acquisition.

2,000+ Total users generated
5.50%+ Weekly users over time
95%+ Enabling rate by users
2 Major browser platforms
rate enabling

One of the most interesting signals was that the extension continued to be used while people were actively using their Pinterest accounts. This suggested that the product was not simply being installed and forgotten, but was becoming part of the browsing workflow for a portion of its users.

07 · LESSONS

What I learned from the project

01 · Start with the user problem

The initial idea did not begin with “I want to build a Chrome extension.” It began with observing a change in the user experience and asking whether users needed more control.

02 · A small product can solve a focused problem

Instead of attempting to compete with Pinterest, the project focused on one specific layer of the existing experience.

03 · Distribution is part of engineering

Building the product was only half of the challenge. Understanding search intent and creating an organic acquisition strategy became an important part of the product itself.

04 · Usage matters more than installation numbers

A browser extension can accumulate installations without becoming useful. Looking at how users actually interact with the extension provides a more meaningful signal of product value.

08 · IMPACT

From an observation to a real product

What started as an observation about the changing nature of Pinterest content became a real software product used by thousands of people.

The project allowed me to work across several areas at once: product discovery, browser-extension development, user experience, SEO, organic acquisition, and product analytics.

More importantly, it reinforced an approach I want to continue applying to future products:

Find a real change in user behavior, identify the friction it creates, build the smallest useful solution, and then find a sustainable way for the people who need it to discover it.

THE TAKEAWAY

Technology is only the beginning.

The Pinterest AI Content Filter project was not just about detecting or filtering AI-generated images. It was an exercise in turning an emerging user problem into a focused product, validating it through real usage, and building an organic acquisition strategy around the problem itself.

The extension continues to evolve as the relationship between AI-generated content and online visual platforms changes.

Built from a problem. Validated by users. Grown organically.

More case studies

  • Rawnaq Al madinahrawnaqalmadinah.com

    Rawnaq Al madinah

    Transforming an Offline Decoration Business into a Search-Driven Digital Presence

  • Observability & Security LayerRequest access

    Observability & Security Layer

    Designing a centralized infrastructure layer to monitor application performance, system health, logs, and security events.

  • Content Automation SystemRequest access

    Content Automation System

    An internal system I built to automate content research, analysis, AI generation, captions, and distribution across Facebook and Pinterest

  • Smart Detector — IoT-Based Gas Leak Detection and Automatic Extractiondrive.google.com

    Smart Detector — IoT-Based Gas Leak Detection and Automatic Extraction

    Developed an Arduino-based gas safety system using C++ to detect butane leaks, trigger visual and remote alerts, and activate a fan and servo-controlled window to help remove accumulated gas.