BOTiqAI

• 

  Integrated Chat Bot for product recommendations to users

• 

  Cart view with added containers

• 

  Product List

• 

  Architecture diagram with added Microservices

Traditional e-commerce platforms often lack intelligent, conversational assistance that can truly understand customer intent and deliver personalized recommendations in real time. Our project reimagines online shopping by integrating AI-driven conversational search with a scalable Kubernetes-native microservices architecture.

Inspiration

We were inspired by the rapid shift in consumer behavior toward AI-first interactions—from chatbots to recommendation engines. Traditional e-commerce sites felt static, with rigid search bars and limited personalization.

We asked ourselves: What if shopping online could feel like talking to a trusted store assistant who knows your tastes, context, and intent?

This question led us to combine Google’s Online Boutique demo (a cloud-native microservices reference app) with Gemini AI to create something truly interactive.

What it does

• Customers can chat naturally with an AI assistant to find products.
  
• Gemini AI provides personalized recommendations in real time, based on intent and preferences.
  
• A cloud-native microservices architecture ensures scalability, resilience, and enterprise readiness.
  

How we built it

1. Foundation – Deployed Google’s Online Boutique reference app on Google Kubernetes Engine (GKE), with 11+ independent microservices.
   
2. AI Layer – Built a custom Gemini AI service to handle:
   
   • Natural language product queries
     
   • Personalized recommendations
     
   • Conversational context tracking
     
3. Frontend Chat UI – Integrated a real-time chat interface where users interact with the AI assistant.
   
4. Cloud-Native Deployment – Containerized all services, orchestrated via Kubernetes, ensuring autoscaling and fault isolation.
   

Challenges we ran into

1. Context Management – Keeping multi-turn conversations relevant to both user preferences and product catalog.
   
2. Deployment – Debugging Kubernetes YAML, IAM roles, and ensuring smooth CI/CD pipelines.
   

Accomplishments that we're proud of

1. Built a conversational shopping assistant that feels natural and engaging.
   
2. Achieved scalable, resilient deployment on GKE, capable of handling production-level workloads.
   
3. Created a solution that bridges AI innovation with enterprise-grade cloud systems.
   

What we learned

1. How to bring LLMs and microservices together to solve real business problems.
   
2. Why good observability (logs, metrics, tracing) is key when debugging a system with many moving parts.
   
3. Ways to make AI responses faster without losing quality.
   
4. That designing with an AI-first mindset can make online shopping feel much more natural and human.
   

What's next for BOTiqAI

1. Multi-modal AI – Letting shoppers search with images (e.g., “show me shoes like this”).
   
2. Smarter personalization – Learning more about each customer’s style and preferences over time.
   
3. Enterprise-ready add-ons – Easy integrations with platforms like Shopify and Magento.
   

Built With

• ai-studio)-containerization-&-orchestration:-docker
• asp.net-core
• c#-(.net-9.0)
• cloud-build
• cloud-profiler
• cloud-trace
• cors-frontend-technologies:-html5
• css3
• docker-compose
• environment-variables
• express.js-ai-&-machine-learning:-google-gemini-ai
• go-templates
• google-cloud-spanner
• google-generative-ai-databases-&-caching:-redis
• gorilla-mux
• helm
• http/rest
• java-frameworks-&-libraries:-flask
• javascript
• json
• kubernetes
• kustomize-communication-&-apis:-grpc
• logrus-(go)
• ngrok
• node.js
• pino-(node.js)-security-&-configuration:-google-cloud-secret-manager
• postgresql-cloud-platform-&-services:-google-cloud-platform-(gke
• programming-languages:-go-(golang)
• protocol-buffers
• python
• responsive-design-development-&-deployment-tools:-skaffold
• secret-manager
• terraform
• xml-observability-&-monitoring:-opentelemetry