AWS Q Business & Salesforce Agentforce Agentic AI Case Study

Situation

As part of my continuous learning in AI and agentic tooling, I explored AWS Q Business and Salesforce Agentforce to develop AI-driven virtual agents. My objective was to build a conversational AI solution capable of answering customer inquiries related to comparing Chevy and Ford trucks using structured and unstructured automotive data.

Task

I designed and implemented an AWS Q Business virtual agent within my AWS environment, enabling customer inquiries to be answered via an AI service by leveraging automotive datasets in the knowledgebase I created.

Actions Taken

1. Set up an AWS environment, establishing an organization, user accounts, groups, and roles via AWS Identity Center.

2. Collected and organized structured and unstructured data from Ford and Chevy websites, including brochures, truck specifications, and pricing details.

3. Configured an S3 bucket to store and manage the collected datasets / knowledgebase.

4. Created and configured an AWS Q Business agent.

5. Established a RAG (Retrieval Augmented Generation) process utilizing the AWS S3 connector to connect and ground the model with uploaded data set / knowledgebase.

6. Synced the data, tested, and refined the AI agent using the AWS web interface.

7. Configured the solution to sync the data one time per week as additional data is added to the knowledgebase.

8. Monitored cost usage via AWS Console and cost management dashboards to ensure compliance with free-tier limits.

Results

• The AWS Q Business agent successfully retrieves and responds to customer inquiries using the uploaded datasets.

• Examples of supported inquiries:

  • “What engine sizes are available for a Chevy Silverado 1500 vs. a Ford F-150?”

  • “What is the average cost of a new Chevy Silverado 3500?”

  • “What interior colors are available for a Ford Raptor?”

  • “Does the Chevy Silverado High Country offer cloth seats?”

Key Takeaways & Lessons Learned

• Hands-on AI Development – Strengthened skills in AWS environment setup, data management, and RAG and AI resource configuration to create functional customer service agents.

• Data Quality is Critical – AI accuracy depends on clean, high-quality, and well-structured data. A broad, unbiased dataset is essential to minimize hallucinations and improve reliability and accuracy.

• Low-Code AI Development – Modern platforms like AWS and Salesforce empower technical business users to build AI-driven virtual agents with minimal coding.

Note, I have setup Salesforce Agentforce agents as well as part of my learning and certification activities.