The Autonomous RFP Bidder
I built this solution using IBM watsonx Orchestrate as the core agentic framework. Foundation Model: I selected the Llama-3-2-90b model for its strong reasoning capabilities. Knowledge Base (RAG): I utilized Orchestrate's Knowledge feature to ingest unstructured PDF data ('Risk Guidelines' and 'Product Catalogs'), allowing the agent to ground its answers in company policy. Agentic Reasoning: I implemented a ReAct (Reason + Act) style architecture. This forces the agent to 'think' in steps: first analyzing the user's prompt, then searching the Knowledge Base for compliance rules (rejecting bids under $10k), and finally performing calculations to generate a proposal. Tools & Content Generation: I leveraged the model's generative capabilities to autonomously draft professional email content directly within the chat interface, ensuring all proposals are ready for review before being sent.
Category tags:
For Innovators & Creators