Improving Accuracy of AI Agents in Oracle Fusion: Prompt Optimization and Training Agents
Summary:
We are currently using multiple LLM-based AI agents within Oracle Fusion use cases, and we’ve observed that over time the accuracy and consistency of responses is gradually decreasing.
We wanted to understand the best practices from the community around improving and maintaining accuracy in such setups. Is there any way we can train our seeded/custom AI Agents through Prompts or with the help of any other tool.
Content (please ensure you mask any confidential information):
Version (include the version you are using, if applicable):
Code Snippet (add any code snippets that support your topic, if applicable):
0