Thiago asked I post here. I have built a streamlit AI agent configuration manager that pushes AI agent configuration to an indepedent streamlit AI chat agent (few lines of code to drop fully baked AI agents into streamlit chat. This configuration manager solves notable weaknesses in incorporating AI agents into businesses, including intent scoping, executing actions, privacy and security awareness, requiring specific data inputs, business API communications, etc… See the pinionai.com website. The manager builder also (depending on permission) allows streamlit AI agent to accept and be serviced by a live human agents in the management app. …functioning like a call center also… BUT When doing this (transferring an AI agent bot to a live agent) the back and forth communication between user and human agent has been some what heavy handed in what was possible by need to perform frequent polling to listen for changes and responses. I am hoping to show you what I’m doing… as it is in fact crazy cool, but also hope to make a case for an async post capability to an agent backend that would trigger front end event activity. Please reach out. I look forward to showing you the power of this streamlit solution, but also what would make it even better. Stack is: Streamlit, Go (for speedy API) and Postgres.
I’d be happy to have a conf call. Sounds like an interesting use case.
Send me a DM.
1 Like
asehmi. I’ve been searching where to send you a private message, and don’t seem to have this ability. Perhaps because I’m new to the boards. If you want drop me a message on this page with your info and I’ll email you direct. Pinion AI | Contact Us Would like to meet with you. I can perhaps meet tomorrow, but I will be out next week, so perhaps we can schedule for a second week in Feb timeframe. Let me know, look forward to meeting with you soon.
1 Like