Hi everyone, let me give you some context about my project. I'm working on a customer service chatbot using DialogflowCX, Cloud Run with Gemini, LangChain, and Vertex AI Search. Currently, I need my chatbot to be more conversational because users are being ambiguous with the products they want in their quotation.
I've experimented with a few approaches. I tried Conversational Agents, but I've had problems getting it to work how I wanted, especially because my product data has a struct with product measurements, and it seemed to behave strangely with the JSONL datastore. I also experimented with LangGraph, but I'm now interested in using ADK instead. This is especially appealing because I need to build tools to calculate measures and unit conversion.
- Is this a suitable use case for ADK? I was told by my Google Support that ADK is only for AgentSpace agents and that I should stick to DFCX.
- Is ADK ready for production usages?
- For someone looking to transition from Dialogflow CX to ADK for more complex agentic behaviors, what's the learning curve like, and are there specific resources you'd recommend?
I'm all ears for any suggestions or guidance you might have based on your experiences!
LinkedIn
Twitter
