Skip to content

Commit 73f85de

Browse files
brown9804brown9804
authored
Refactor README for AI Shopping Assistant and analysis
Updated the README to improve structure and clarity, adding sections for AI Shopping Assistant and customer behavior analysis. Enhanced descriptions and added images for better visualization.
1 parent d8a972c commit 73f85de

1 file changed

Lines changed: 40 additions & 20 deletions

File tree

  • 0_Azure/_industry-specific/1_Retail/0_Art_of_the_possible

‎0_Azure/_industry-specific/1_Retail/0_Art_of_the_possible/README.md‎

Lines changed: 40 additions & 20 deletions
Original file line numberDiff line numberDiff line change
@@ -10,32 +10,52 @@ Last updated: 2025-10-23
1010

1111
----------
1212

13-
1. AI Shopping assistant: `my own copilot`, `want to chat with my data` [GPT-RAG Solution Accelerator](https://github.com/Azure/GPT-RAG). RAG-based chatbot to interact with manuals, guides, and product data for better search and personalization.
13+
## AI Shopping assistant
1414

15-
| [Public Architecture](#basic-architecture) | [Private Architecture](./1_PrivateArchitecture.md) | [Standard Zero-Trust Architecture](https://github.com/Azure/GPT-RAG) |
16-
| --- | --- | --- |
17-
| <img width="800" alt="image" src="https://github.com/user-attachments/assets/5444e87c-32af-44e8-aa86-22fe4082c4f4" /> | <img width="800" alt="image" src="https://github.com/user-attachments/assets/af835493-2d76-4216-8df6-3e258c9db949"> | <img width="800" alt="image" src="https://github.com/user-attachments/assets/67e7d9eb-f757-4fe9-85f4-d5bb08b9e55f" /> |
18-
19-
2. Analyze customer behavior and feedback for trend discovery. [Demo: How to configure the Microsoft Fabric (Power BI) MCP connector in Copilot Studio](https://github.com/MicrosoftCloudEssentials-LearningHub/Fabric-MCP-Agent2Agent)
15+
> `my own copilot`, `want to chat with my data` [GPT-RAG Solution Accelerator](https://github.com/Azure/GPT-RAG). RAG-based chatbot to interact with manuals, guides, and product data for better search and personalization.
2016
21-
> Low code approach (Copilot Studio Agents):
22-
23-
<img width="1141" height="669" alt="MCP-Fabric-Agent2Agent-CopilotStudio-databricks+data-agents+copilot-Studio drawio" src="https://github.com/user-attachments/assets/76be59e3-f48e-4c01-aa58-464463658d92" />
24-
25-
> Full code approach (AI Foundry Agents):
26-
27-
<img width="1141" height="668" alt="MCP-Fabric-Agent2Agent-CopilotStudio-data agent Fabric + AI Foundry drawio" src="https://github.com/user-attachments/assets/5742dc3a-427d-4ab5-862d-ba291a31fc35" />
17+
| [Public Architecture](#basic-architecture) | [Private Architecture](./1_PrivateArchitecture.md) | [Standard Zero-Trust Architecture](https://github.com/Azure/GPT-RAG) |
18+
| --- | --- | --- |
19+
| <img width="800" alt="image" src="https://github.com/user-attachments/assets/5444e87c-32af-44e8-aa86-22fe4082c4f4" /> | <img width="800" alt="image" src="https://github.com/user-attachments/assets/af835493-2d76-4216-8df6-3e258c9db949"> | <img width="800" alt="image" src="https://github.com/user-attachments/assets/67e7d9eb-f757-4fe9-85f4-d5bb08b9e55f" /> |
2820

29-
3. Customer360 AI Experience:
30-
- Enable AI-driven suggestions for complementary products
31-
- Real-time sentiment analysis and campaign triggers based on trends.
32-
- Improve data taxonomy and metadata
33-
- Enhance user experience with personalized content blocks.
21+
## Analyze customer behavior and feedback for trend discovery
3422

35-
<img width="1656" height="921" alt="retail-ai-samples drawio" src="https://github.com/user-attachments/assets/c7b41c55-fd5e-4e6f-af1c-d206e32e0d51" />
23+
[Demo: How to configure the Microsoft Fabric (Power BI) MCP connector in Copilot Studio](https://github.com/MicrosoftCloudEssentials-LearningHub/Fabric-MCP-Agent2Agent)
3624

37-
25+
> Low code approach (Copilot Studio Agents):
3826
27+
<img width="1141" height="669" alt="MCP-Fabric-Agent2Agent-CopilotStudio-databricks+data-agents+copilot-Studio drawio" src="https://github.com/user-attachments/assets/76be59e3-f48e-4c01-aa58-464463658d92" />
28+
29+
> Full code approach (AI Foundry Agents):
30+
31+
<img width="1141" height="668" alt="MCP-Fabric-Agent2Agent-CopilotStudio-data agent Fabric + AI Foundry drawio" src="https://github.com/user-attachments/assets/5742dc3a-427d-4ab5-862d-ba291a31fc35" />
32+
33+
## Customer360 AI Experience
34+
35+
- Enable AI-driven suggestions for complementary products
36+
- Real-time sentiment analysis and campaign triggers based on trends.
37+
- Improve data taxonomy and metadata
38+
- Enhance user experience with personalized content blocks.
39+
40+
<img width="1656" height="921" alt="retail-ai-samples drawio" src="https://github.com/user-attachments/assets/c7b41c55-fd5e-4e6f-af1c-d206e32e0d51" />
41+
42+
> [!TIP]
43+
> 1. **Complementary Product Suggestions**: Gold layer → AI Foundry/OpenAI → Recommendation API → Web App.
44+
> 2. **Real-Time Sentiment & Campaign Triggers**: Eventstream → KQL → Data Activator → Marketing automation.
45+
> 3. **Improve Taxonomy & Metadata**: Purview ensures consistent metadata and compliance across layers.
46+
> 4. **Personalized Content Blocks**: Azure ML models → App Service → Front Door → End-user experience.
47+
48+
| **Layer / Component** | **Details** |
49+
|------------------------|-------------|
50+
| **1. Data Sources** | • **E-commerce Platform**: Provides product catalog, pricing, inventory, and order history.<br>• **Customer Interaction Data**: Includes clickstream, browsing patterns, and purchase behavior.<br>• **Social Media APIs (X, YouTube, Instagram, TikTok)**: Streams hashtags, influencer posts, and comments for trend and sentiment analysis.<br>• **API Connections**: Enable secure data pulls from external sources into the platform. |
51+
| **2. Ingestion Layer** | • **Fabric Data Factory**:<br> – Handles **batch ingestion** of structured data from e-commerce and customer systems.<br> – Moves raw data into the **Bronze layer** in OneLake.<br>• **Eventstream (Fabric)**:<br> – Captures **real-time social media feeds** and customer interaction events.<br> – Streams data directly into OneLake and **Real-Time Analytics (KQL)** for immediate processing. |
52+
| **3. Medallion Architecture in Fabric** | • **Bronze (Raw)**: Stores unprocessed data from APIs and batch pipelines.<br>• **Silver (Cleaned/Transformed)**:<br> – PySpark notebooks clean, standardize, and enrich data.<br> – Apply taxonomy rules and metadata alignment.<br>• **Gold (Curated)**:<br> – Business-ready tables for analytics and AI models.<br> – Includes aggregated sentiment scores, product attributes, and customer segments. |
53+
| **4. Governance & Metadata (Microsoft Purview)** | • Scans Silver and Gold layers for lineage and compliance.<br>• Maintains taxonomy dictionaries and metadata standards.<br>• Applies sensitivity labels and access policies for secure data usage. |
54+
| **5. Real-Time Analytics** | • **KQL Database (Fabric Real-Time Analytics)**:<br> – Processes Eventstream data for sentiment scoring and trend detection.<br> – Outputs aggregated insights to Gold layer for reporting and AI consumption.<br>• **Data Activator / Logic Apps**: Triggers marketing campaigns or feature flags when sentiment thresholds are met. |
55+
| **6. AI & Machine Learning** | • **Azure AI Foundry + OpenAI**:<br> – Generates complementary product suggestions using:<br>  • Content-based filtering (attributes, taxonomy).<br>  • RAG (Retrieval-Augmented Generation) for product guides.<br>• **Azure Machine Learning**:<br> – Trains personalization models (propensity, uplift).<br> – Deploys endpoints for real-time scoring.<br>• **AI Vision & AI Services**: Optional image analysis for social media content. |
56+
| **7. Delivery Layer** | • **Container App / Web App**: Hosts APIs for personalized content blocks and recommendation engine.<br>• **Azure Front Door/CDN**: Ensures global, low-latency delivery of personalized experiences.<br>• **App Service Plan**: Provides scalable hosting for web applications. |
57+
| **8. Insights & Reporting** | • **Power BI (Direct Lake Mode)**:<br> – Connects to Lakehouse in OneLake for real-time dashboards.<br> – Displays campaign performance, sentiment trends, and recommendation effectiveness.<br>• **Copilot**: Assists business users with natural language queries on data. |
58+
| **9. Security & Monitoring** | • **Azure Key Vault**: Stores API keys, credentials, and secrets securely.<br>• **Azure Monitor & Application Insights**: Tracks pipeline health, model latency, and API performance. |
3959

4060
<!-- START BADGE -->
4161
<div align="center">

0 commit comments

Comments
 (0)