In this series, we’ll build an AI-powered Product Recommendation Engine using Laravel + GPT, where suggestions are generated based on user behavior and product metadata using function calling and optionally embeddings.
⚙️ Step 1: Create a fresh Laravel project
composer create-project laravel/laravel gpt-recommender
cd gpt-recommender
🛒 Step 2: Create Product and User models
php artisan make:model Product -m
php artisan make:model User -m
php artisan make:migration create_user_product_views_table
Update the migrations:
create_products_table.php:
public function up()
{
Schema::create('products', function (Blueprint $table) {
$table->id();
$table->string('name');
$table->text('description')->nullable();
$table->decimal('price', 10, 2);
$table->string('category')->nullable();
$table->timestamps();
});
}
create_users_table.php (if not default):
Ensure it has at least:
$table->string('name');
$table->string('email')->unique();
create_user_product_views_table.php:
public function up()
{
Schema::create('user_product_views', function (Blueprint $table) {
$table->id();
$table->foreignId('user_id')->constrained()->onDelete('cascade');
$table->foreignId('product_id')->constrained()->onDelete('cascade');
$table->timestamp('viewed_at')->default(now());
});
}
Run migrations:
php artisan migrate
📦 Step 3: Seed some dummy products
php artisan make:seeder ProductSeeder
In ProductSeeder.php:
use App\Models\Product;
public function run(): void
{
$products = [
['name' => 'Wireless Headphones', 'description' => 'Bluetooth, noise-cancelling', 'price' => 299.00, 'category' => 'Electronics'],
['name' => 'Ergonomic Chair', 'description' => 'For office comfort', 'price' => 499.00, 'category' => 'Furniture'],
['name' => 'Smartwatch', 'description' => 'Fitness tracker & notifications', 'price' => 399.00, 'category' => 'Electronics'],
['name' => 'Standing Desk', 'description' => 'Adjustable height, wood top', 'price' => 999.00, 'category' => 'Furniture'],
];
foreach ($products as $product) {
Product::create($product);
}
}
In DatabaseSeeder.php:
$this->call(ProductSeeder::class);
Seed the database:
php artisan db:seed
✅ Output
You now have:
- A
productstable with items to recommend - A
user_product_viewstable to track interest - The basic structure for connecting user behavior to GPT
✅ Up next (Day 2): we’ll track user views, then use GPT function calling to suggest related products based on their view history.
