@Iderah Expert The answer is inside…ask yourself what you care the most about…something that you’ve been doing over the last year without even thinking about it too much…what interest you online etc. The business is you…so start there

@Aziz Diallo Thats the reason I’m not getting into outbound voice bots. Here is the info directly from the source (FCC in USA): https://www.fcc.gov/document/fcc-makes-ai-generated-voices-robocalls-illegal But with that said there are AI vendors that found a way to work around it, at least until new legislation will be added/updated. https://aisdr.com/platform/

@Viktor Van der wolf You have to take the time to understand the type of data that is being queried to determine to best approach. Based on your question it sounds like you have very limited experience with this. Let me give you couple of considerations, is the type of responses needed from you agent is around an order status from Monday that may include order details etc, then there is no DB here. This would be an API call since data in Monday is going to be dynamic and you are not going to replicated live another DB synced with Monday (which is a SaaS anyway). The second consideration, you want to build a RAG? for training data for your agent to use for its responses? Then your question starts to make more sense. You can place your RAG on any number of database types. But before storing data into your vector database you need to chunk it based on the data structure and test it so you get accurate results when your agent queries the vector db. Makes sense ? My advice here is to take the time to design the right solution for your customer. Not all AI agents are equal in terms of architecture and build. Good luck.

How about…trust the person…then the process…people->process->tools