A simple example implementation of the VoiceRAG pattern to power interactive voice generative AI experiences using RAG with Azure AI Search and Azure OpenAI's gpt-4o-realtime-preview model.
I've encountered an issue where the system_message is successfully transmitted to the server, but it doesn't seem to be affecting the AI's behavior as expected. The message is correctly set in the RTMiddleTier instance and appears to be included in the session update. However, the AI's responses do not reflect the instructions provided in the system message.
I've encountered an issue where the system_message is successfully transmitted to the server, but it doesn't seem to be affecting the AI's behavior as expected. The message is correctly set in the RTMiddleTier instance and appears to be included in the session update. However, the AI's responses do not reflect the instructions provided in the system message.
How can I resolve this issue?