Generative AI Customer Example: Time To Say Dubai – Chatbot

1. Business Problem

Dubai is a global hub for tourism, business, education, and employment, attracting millions of visitors and expatriates annually. However, finding reliable and comprehensive information about Dubai can be challenging, as users must navigate multiple websites for details on visas, laws, schooling, jobs, accommodation, and tourism. Existing solutions are fragmented, requiring users to piece together information from different sources.

2. Project Objective

To create an AI-powered, multilingual chatbot that provides comprehensive information about Dubai, covering topics such as tourism, schooling, accommodation, visa processes, and laws. Additionally, the chatbot integrates a job search functionality that allows users to find relevant job opportunities in the UAE based on their position and location. It supports English and German through intelligent language detection and translation.

3. Business Problem and Traditional Approach

Traditional Approach Challenges:

• Manual Information Search: Users had to navigate multiple websites, government portals, and recruitment agencies to find relevant information.
• Delayed Customer Support: Businesses and agencies had slow response times due to manual interactions.
• Fragmented Information Sources: Users struggled with inconsistent or outdated data.
• Limited Job Search Efficiency: Traditional platforms lacked conversational assistance and personalization.
• Language and Accessibility Barriers: Non-native speakers faced difficulties understanding complex procedures.

4. Generative AI Solution on AWS

Solution Architecture

The AI-powered chatbot leverages AWS services to enhance automation, multilingual capabilities, and user experience. The architecture follows a serverless design to ensure scalability and cost optimization.

• Frontend Deployment: Deployed using AWS Amplify for secure, scalable web hosting.
• Apify Webhook: Entry point for receiving job search webhook events.
• API Gateway: Manages incoming webhook and chatbot requests.
• AWS Lambda (Function 1): Processes incoming webhook requests from Apify.
• Amazon SNS: Passes messages between Lambda functions.
• AWS Lambda (Function 2): Interacts with RDS and triggers workflows.
• Amazon RDS: Stores processed job data.
• Amazon Comprehend: Detects user input language.
• Amazon Translate: Translates bot responses and job listings to the user’s preferred language (English or German).
• Amazon Bedrock (Claude Model): Generates intelligent responses.
• Amazon Lex: Enables conversational AI capabilities.

AWS Generative AI-Specific Services Used

• Amazon Bedrock (Claude Model): Provides dynamic, natural language responses.
• Amazon Lex: Enables text-based natural conversation.
• AWS Lambda: Facilitates serverless processing.
• Amazon Comprehend: Detects input language in real time.
• Amazon Translate: Ensures accurate multilingual translations for responses and job listings.

5. Quantitative Business Impact Metrics

Metric	Before AI Implementation	After AI Implementation
User Query Response Time	~5–10 minutes (manual search)	~2 seconds (AI-powered bot)
Customer Support Cost	High (manual agents required)	Reduced by 60% (automation)
Job Search Success Rate	70% match efficiency	85% match efficiency
User Engagement	5–7 minutes per query	2–3 minutes (bot-driven)

6. Business-Led Transformation Using Generative AI

• Automated Customer Assistance: Users access information instantly using natural language queries.
• Multilingual Support: Real-time detection and translation in English and German using Amazon Comprehend and Translate.
• Intelligent Job Matching: Personalized recommendations based on role, experience, and location.
• Personalized Experiences: AI tailors responses to user preferences.

7. Deployment and Production Readiness on AWS

• Deployment Environment: AWS Cloud
• Frontend Hosting: AWS Amplify
• Production Status: Fully deployed and operational
• Scaling Strategy: Auto-scaling via Lambda and Lex
• Security Measures: IAM, API Gateway auth, AWS WAF, KMS

8. Project Documentation and Deliverables

• Evidence of AWS Generative AI Practice Implementation: Statement of Work (SOW)

9. Conclusion

The AI-powered chatbot for Time To Say Dubai has revolutionized how users access information and job opportunities in Dubai. Leveraging Amazon Bedrock, Amazon Lex, Amazon Comprehend, and Amazon Translate, the chatbot offers multilingual, scalable, and intelligent support with significantly improved engagement and reduced costs. This successful deployment exemplifies the transformative potential of Generative AI on AWS.

Solution Architecture: