If an agent task takes 2 minutes to complete, how do you architect the API so the user's browser connection doesn't time out?

AI agents often rely on external tools/APIs. How would you design a system that ensures robustness when these dependencies are unreliable or slow?

If a tool returns a massive 50MB JSON file, you can't feed that to the LLM. How do you "summarize" or "filter" tool observations for the agent?

If your user queries are short "slang" phrases but your documents are formal legal texts, how do you ensure the embedding model can bridge that semantic gap?