Forem: KAOUTAR BENHADINE

Bot or Not: Can You Spot the AI? 🎨

KAOUTAR BENHADINE — Mon, 29 Sep 2025 06:37:46 +0000

This is a submission for the KendoReact Free Components Challenge.

TL;DR

Built an AI image detection game using Kiro AI IDE + KendoReact MCP Server to generate components & override styles effortlessly—turning traditionally tedious UI customization into a breeze! ⚡

What I Built

Bot or Not challenges players to distinguish AI-generated images from real photos—a critical skill in today's AI-powered world! 🌍

Two Game Modes:

🎯 Daily Challenge: 3 rounds with results reveal
🔥 Streak Mode: Unlimited rounds with progressive difficulty
⚙️ Admin Panel: Manage images, create pairs, monitor stats

Why it matters:

Sharpen your ability to spot AI-generated content—an increasingly valuable skill!

Demo

🎮 Live Demo

📦 GitHub

🎥 Video Demo:

KendoReact Components Used

+10 KendoReact packages powering the entire experience:

Core UI Components:

🔘 Buttons - Game actions, navigation, mode selection
🎴 Layout (Card, PanelBar, StackLayout) - Image containers, game structure
📋 Grid - Admin panel data management
💬 Dialogs (Dialog, Window) - Confirmations, instructions, results screens

Form & Input Components:

📝 Dropdowns (DropDownList, ComboBox) - Category & difficulty filters
📤 Upload - Admin image uploading
🔤 Inputs (TextBox, NumericTextBox, Switch) - Admin forms & toggles
🏷️ Labels - Form labels & accessibility
📋 Form - Structured admin input handling

Feedback & Visualization:

🏅 Indicators (Loader, Badge) - Loading states, scores, streaks
📊 ProgressBars - Round progress (1/3, 2/3, 3/3)
🔔 Notification - Success messages, alerts, feedback

Additional:

🎨 Kendo Theme Default - Consistent styling
🎯 SVG Icons - Visual polish throughout

AI Coding Assistant Usage

What is Kiro? 🤔
Kiro AI IDE is an AI-powered development environment that integrates with MCP servers to provide intelligent coding assistance.

The MCP Server Magic: Two Game-Changing Features ✨

1️⃣ Instant Component Generation 🏗️

Without AI Coding Assistant Usage: Hours writing boilerplate, reading docs, configuring props

With AI Coding Assistant Usage: Describe what you need → Get production-ready components

Me: "Add a progress bar to the daily challenge session" Kiro + KendoReact MCP Server: *Complete Progress Bar with proper config, TypeScript, accessibility*

🎥: Add KendoReact UI Component & Fix Bugs

2️⃣ Effortless Style Overrides 🎨 - THE GAME CHANGER!

The Traditional Nightmare: 😫

Inspect nested CSS classes
Trial-and-error with specificity
Break component functionality
Spend hours on simple visual tweaks

With AI Coding Assistant Usage: ⚡

Me: "Make Group Button have mint green gradient backgrounds with hover animation" Kiro + KendoReact MCP Server: *Perfect CSS overrides, maintains functionality*

🎥: Add Custom Style To KendoReact UI Component

This alone saved 200% of UI development time! 🎯
Real Impact 💪

Developer Experience:

❌ No more CSS frustration
❌ No more documentation rabbit holes
✅ Focus on creative design

Key Takeaway:
The KendoReact MCP Server isn't just convenient—it fundamentally changes how you work with component libraries. What used to be frustrating (style overrides) became the easiest part! 💎

Happy prompting⚡

Global Sanctions Index

KAOUTAR BENHADINE — Mon, 28 Jul 2025 06:55:03 +0000

This is a submission for the Algolia MCP Server Challenge

What I Built

My inspiration comes from spending 20 working days implementing OFAC sanctions screening for insurance clients - a problem now solvable in under 2 hours with Algolia MCP Server and Claude Desktop.

My submission is a Global Sanctions Index that consolidates international sanctions data from US OFAC, EU, and UN sources into a unified, AI-optimized index using Algolia MCP Server and Claude Desktop.

Core Innovation:

🔄 Multi-Source Consolidation: Automated processing of OFAC SDN/Non-SDN, EU Financial Sanctions Files, and UN Security Council lists

🧠 AI-Powered Deduplication: Claude Desktop identifies duplicate entities across different international naming conventions and data formats

🎯 Real-Time Optimization: Direct Claude-Algolia integration for dynamic index configuration and performance tuning

Real-World Impact

This addresses genuine compliance pain points that financial institutions face daily:

Manual screening delays
High false positive rates
Inconsistent international data
Regulatory compliance gaps

The Claude Desktop + Algolia MCP combination solves these challenges while maintaining sub-100ms performance for real-time transaction processing.

Demo

How I Utilized the Algolia MCP Server

This implementation integrates Claude Desktop directly with Algolia MCP Server for advanced backend data processing. Claude Desktop handles the complex AI analysis while Algolia MCP Server manages search optimization and index performance in real-time coordination.

🤖 NO CODE, JUST PROMPTS!

- 🔍 Data Harmonization via Claude Desktop

Analyze sanctions data from OFAC CSV and EU XML formats. Normalize entity names, addresses, and identification numbers into unified schema while preserving source attribution.

- 🧠 AI-Powered Entity Resolution

Compare these potential duplicate entities considering transliteration variations (Vladimir/Владимир), cultural naming patterns, and corporate relationships. Provide confidence-scored merge recommendations.

- ⚡ Dynamic Index Optimization

Based on current query patterns and data distribution, recommend optimal faceting strategy, ranking algorithms, and search synonyms for international sanctions screening.

Key Takeaways

The Core Problem

International sanctions data exists in silos across multiple jurisdictions - US OFAC, EU, and UN each maintain different data structures and standards, updating on independent schedules with varying naming conventions.

This fragmentation creates massive compliance gaps where the same sanctioned entity appears differently across lists, making traditional string matching ineffective for international names and corporate structures.

The Breakthrough

This solution leverages semantic search through Claude Desktop's intelligence combined with Algolia MCP Server's optimization to understand entity relationships, name variations, and transliterations across languages, enabling unified search that accurately identifies entities while avoiding false positives.

This approach solved the core challenge of handling entity deduplication across different languages, legal systems, and naming conventions where traditional methods fail.

Sample queries

Semantic Search :

Show me Russian oligarchs sanctionned after the Ukraine invasion.

Real-time transaction screening :

Screen a wire transfer from "Vladimir Petrov" (Russia) to "Iranian Shipping Co" (Iran) for $250,000. Show comprehensive sanctions checking with detailed compliance recommendations.

Entity Resolution Showcase:

Show how "PUTIN, Vladimir Vladimirovich" (OFAC), "Vladimir PUTIN" (EU), and "Владимир Путин" (Cyrillic) are identified as the same entity with 95%+ confidence scoring.

Conclusion

The Global Sanctions Index demonstrates how Algolia's MCP Server can orchestrate sophisticated AI workflows for real-world compliance challenges. By combining Claude Desktop's semantic intelligence with Algolia's search performance, the solution delivers both technical excellence and practical business value.

Happy building! 🚀🧠

🤖 Autonomous Sales Intelligence Agent: Zero-Touch Email-to-Sale Automation

KAOUTAR BENHADINE — Mon, 07 Jul 2025 06:24:54 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

🚀 What I Built

I created an Autonomous Sales Intelligence Agent that transforms raw customer emails into structured sales data and automated responses - completely hands-free. This AI-powered system eliminates the manual grunt work of processing customer inquiries by automatically:

🔍 Detecting sales opportunities from incoming emails
📊 Extracting structured data (customer info, products, quantities, pricing)
📋 Generating order records in Google Sheets with unique tracking IDs
📧 Sending personalized confirmation emails to customers
🔔 Notifying sales teams via Slack in real-time
📈 Providing predictive analytics for inventory planning

The system solves the critical problem of sales lead leakage 🚨 - where potential customers slip through the cracks due to delayed responses or missed emails in busy inboxes.

🎥 Demo

🔍 Key Workflow Components Covered:

The Autonomous Sales Intelligence Agent operates through a sophisticated multi-stage workflow that transforms customer emails into actionable business intelligence.

🚀 Initial Setup:

Store credentials and operational mode selection
OAuth integrations for Gmail, Google Sheets, and Slack

📧 Email Processing Pipeline:

Mock data generation for testing
Zapier integration for automatic email capture
DataHub sheet population

🤖 Intelligent Analysis:

Purchase intent detection
Structured data extraction
Unique order ID generation (KB-RunnerH-XXXXXX format)

📊 Multi-Channel Output:

Google Sheets order tracking
Real-time Slack notifications
Automated customer confirmations

🛠️ Advanced Modes:

System Maintenance for data integrity
Predictive analytics for inventory planning

💡 Business Value:

Zero-touch automation
Human-like communication
Seconds vs. hours response time

⚙️ How I Used Runner H

My prompt:

You are my autonomous Sales Intelligence Agent.
Your role is to transform every email into structured sales data and automated customer responses using Gmail, Google Sheets and Slack. 

The task involves setting up a sales intelligence system using Zapier, Google Sheets, Gmail, and Slack. 

The task is complex as it involves multiple integrations and setup processes.

I need zero-hallucination, data-driven sales automation.

🚀 Initial Setup:

Step 0 :

1. Prompt the user for his **Store name** and **email address**.  
2. Obtain OAuth approval to **manage Google Sheet**, **send Gmail** and **send slack message** on his behalf.
3. ask the user to select one of the following operational modes: 
- Batch Email Analysis Mode
- System Maintenance Mode
- Predictive Mode

Step 1 : Preparing mock data
send 5 mails to **email address** where u acte as a customer that wanna make a purchase on his store , the mail should include name of products , there quantity and contact info like physical address , mails should humanitize

Step 2 : Zapier Action processing

ask the user to confirm that zapier has complete his job .it tools about 2min. the Zapier action will add the new mails in the"Sales Intelligence Master" spreadsheet "DataHub" sheet otherwise the sheet will be empty

Step 3:according to the selected mode process one of the following operational modes:

mode 1 : "Batch Email Analysis Mode" :

step 1 : 🔍Mail Scraping :
read "Sales Intelligence Master" (ID : 1gENZ9isErDgNGzDP_NMs0thX0zB-O9TqXadEvkrwdUk ) spreadsheet "DataHub" sheet rows especially the content column to detect potential customers mails where they request products the mail addrees of the customer is included in the "Customer" column

📊 Simple Output Format:
📈 After Each Email Processed:
✅ NEW LEAD DETECTED:
Customer: John Smith (john@company.com)
Product: Industrial Printer (5 units needed)
Total Price: $2000 

step 2: 📋 Fill the "Sales Intelligence Master" spreadsheet "Orders" sheet :
if there is no column header in the sheet put this list of headers OrderId, Customer, Price, DTCreation,DTShipping, DTDelivery, ContactInfo, Details
extract data for each mail content and insert it in the sheet ; 
- generate an order ID in this format 'KB-RunnerH-DDDDDD', the D is for Digets,
- ContactInfo should include the customer mail adress which is equivalent to the "Customer" column in "DataHub" sheet and any relevant data the customer mentioned in his mail "Content" without halucination! sensitive data,
-the detail column  sheet contains the list of product the customer ask for and its quantity.
respect the format of the "Orders" sheet it is as follow : OrderId, Customer, Price, DTCreation,DTShipping, DTDelivery, ContactInfo, Details
Return to the "DataHub" sheet and set the "Status" to "PROCESSED" for each well processed entry 

step 3: 📊 Interact with Slack:
add a message to the slack channel #sales-deals, the message is in this format 'order ID', 'dateCreation','details'; u can update the msg fomat in a better format if needed.

step 4:  send a response mail to the customers using this template at there mail ; the address is in the "DataHub" sheet "Customer" column :

Subject: Re: Order Confirmation - **OrderID**

Hi [Customer_Name],

Thank you for your purchase from **Store name** ! We're thrilled to confirm your order details as follows:

- Order ID: **Order ID**
- Products :**Details**
- Total Price: **Price**

Your order will be shipped to the address you provided: **ContactInfo**.

Should you have any questions or need assistance, feel free to reach out.

Warm regards,
[Store name] Customer Support
the mail should be sent to the same value in the "DataHub" sheet Customer column

step 5: confirm with the user if he want start a new mode
mode 3 : 🛠️ System Maintenance Mode:
step 1 :Clean duplicate entries in Google Sheets
step 2 :Validate data accuracy and completeness
step 3 : Generate system health reports
step 4: send the report to slack channel #sales-deals

mode : 🛠️ Predictive Mode:
where u predict the sales of the folloqing 3 weeks based on the orders tracking in the googlr sheet so the user could prepare his stock and u can add any relevant prediction u find

Always be proactive: after each response, ask the user of which mode to execute unless he write exit

The core innovation lies in the multi-modal operational system powered by Runner H's advanced prompting capabilities:

🧠 Intelligent Email Analysis

Uses natural language processing to identify purchase intent
Extracts structured data without hallucination
Maintains data integrity across multiple touchpoints

🎯 Multi-Mode Operations

📊 Batch Email Analysis Mode: Processes multiple emails simultaneously
🛠️ System Maintenance Mode: Cleans data and validates accuracy
🔮 Predictive Mode: Forecasts sales trends for inventory planning

🤖 Automated Workflows

Dynamic order ID generation (KB-RunnerH-XXXXXX format)
Template-based customer communication
Real-time team notifications

💼 Use Case & Impact

🎯 Primary Beneficiaries:

🛒 E-commerce Store Owners: Eliminate manual order processing
🏢 Small Business Owners: Scale customer service without hiring
💰 Sales Teams: Focus on closing deals, not data entry
📞 Customer Service Reps: Reduce response times from hours to minutes

📈 Business Impact:

⚡ 95% reduction in manual email processing time
🕐 24/7 automated customer response system
🎯 Zero sales lead loss through systematic tracking
📊 Predictive inventory management prevents stockouts
😊 Improved customer satisfaction with instant confirmations

💡 Innovation Differentiators:

🎯 Zero-hallucination data extraction ensures accuracy
🔄 Multi-modal operations adapt to different business needs
🔮 Predictive analytics provide strategic insights
🔗 Seamless integration with existing business tools

🌍 Real-World Applications:

📦 Dropshipping businesses managing high email volumes
🏭 B2B suppliers processing bulk order inquiries
🔧 Service providers converting emails to project requests
🏪 Retail stores automating customer order confirmations

This solution transforms the traditional reactive email management into a proactive sales intelligence system 🔄, turning every customer email into a structured business opportunity while maintaining the human touch through personalized automated responses.

🎯 Conclusion

The Autonomous Sales Intelligence Agent represents a paradigm shift from reactive email management to proactive sales intelligence. By seamlessly integrating AI-powered email analysis with real-time automation across Gmail, Google Sheets, and Slack, this system doesn't just process emails—it transforms every customer interaction into a structured business opportunity.

Happy reading!📚✨