Unlock AI Power in Laravel: Real‑World SDK Use Cases & Code Samples

Overview

The Laravel team recently released the Laravel AI SDK, which provides a clean, intuitive API for integrating AI models and services such as OpenAI, Anthropic, Gemini, Mistral, and Ollama into Laravel applications. The SDK abstracts away model‑specific complexities, enabling natural‑language processing, image recognition, and other AI‑driven features with minimal code.

1. Extracting Insights from User Data

This is the simplest use‑case: an AccountAssistant agent receives a User and Order Eloquent model as context and answers natural‑language questions about the account and recent orders.

namespace App\\Ai\\Agents;

use App\\Models\\User;
use App\\Models\\Order;
use Laravel\\Ai\\Contracts\\Agent;
use Laravel\\Ai\\Contracts\\Conversational;
use Laravel\\Ai\\Promptable;
use Stringable;

class AccountAssistant implements Agent, Conversational
{
    use Promptable;

    public function __construct(public User $user) {}

    public function instructions(): Stringable|string
    {
        return 'You are an assistant helping a customer with their account and recent orders.';
    }

    public function messages(): iterable
    {
        $orders = Order::where('user_id', $this->user->id)
            ->latest()
            ->limit(5)
            ->get()
            ->map(fn ($order) => "- #{$order->id} ({$order->status}) total {$order->total}");

        return [
            new \Laravel\Ai\Messages\Message(
                'system',
                "Here are the user's 5 most recent orders:
" . $orders->implode("
")
            ),
        ];
    }
}

Usage is straightforward:

$assistant = \App\\Ai\\Agents\\AccountAssistant::make(user: $user);
$response = $assistant->prompt('Where is my last order?');
echo (string) $response;

The AI replies with the location of the most recent order based on the provided context.

2. Automated PR Code‑Review Bot

The SDK can power an internal AI code reviewer that comments on Laravel monorepo pull requests.

Typical workflow:

When a PR is created or updated, a CI job or lightweight Laravel service captures the diff and changed files.

Define a CodeReviewer agent whose instructions() method encodes the team’s coding standards (style, architecture, security, performance, etc.).

In prompt(), pass the PR description, diff, and files as attachments, e.g., Files\Document::fromString($diff, 'text/plain').

The agent returns structured output containing: summary: overall change description issues: list of problems with severity, file, and line number suggestions: refactoring tips or missing tests

Finally, use the GitHub/GitLab API to post the results as real comments on the PR.

Key advantage: agents, attachments, and structured output let you enforce team standards automatically.

3. Personalized E‑Learning & Adaptive Quiz System

For an online course platform, the SDK enables intelligent, personalized quizzes.

A. Generate Custom Quizzes

After a student finishes a lesson, invoke a QuizGenerator agent.

Prompt: “You are an exam coach. Generate questions only from the provided course content and tailor them to the student’s weak points.”

Context includes the course summary and the student’s historical quiz results.

Optionally attach course notes or PPT via Files\Document::fromStorage(...).

The agent returns a structured list of questions, which are stored in quizzes and quiz_questions tables.

B. Adaptive Difficulty via Conversation Memory

Implement Conversational or RemembersConversations so the agent remembers previous answers, weak topics, and feedback.

Subsequent quizzes automatically focus on weak areas or gradually increase difficulty.

C. On‑Demand Answer Explanations

A second agent, AnswerExplainer, receives a student's wrong answer and the original question, then returns a customized explanation (e.g., “Why B is incorrect and C is correct”).

D. Enhance with Embeddings & Reranking

Use Embeddings plus whereVectorSimilarTo to locate the most relevant course passages for question generation.

Apply Reranking to reorder candidate questions, ensuring the most appropriate ones appear first.

The entire system is built from Laravel AI SDK primitives: agents, structured output, attachments, conversation memory, embeddings, and reranking.

Summary & Takeaways

The Laravel AI SDK dramatically lowers the barrier to integrating AI into Laravel projects. It supports multiple providers and offers advanced capabilities such as conversation memory, middleware, structured output, attachments, embeddings, and reranking. Real‑world scenarios include user data analysis, automated code review, and personalized education, all achievable with familiar Laravel patterns.

Reference URLs: https://www.amitmerchant.com/real-world-exmples-using-laravel-api-sdk/ https://github.com/laravel/ai https://laravel.com/docs/12.x/ai-sdk