Optimizing Dust: Using Multiple Agents for Better Data Retrieval

Good morning all, I have integrated all my data sources into Dust but the default agent is having a hard time retrieving specific files. I think I may have too much data connected which the platform flagged as having less accurate responses. Does it make sense to setup multiple agents with more specific memory sets (i.e. an agent for meeting transcripts, one for my email archive, etc) and then create one over-arching agent that has more generalized knowledge and can ask specific agents for items?