
I run a cybersecurity advisory firm, and I’ve been evaluating CustomGPT.AI since July. We’re still building and testing, so I wouldn’t call us fully deployed yet. That said, I’ve already built an advisory agent that works from our own material, a tech desk agent, and a multi-agent setup I call “Ask Ryan” that routes questions across agents.
The multi-agent functionality was easier to get working than I expected. Almost everything I do goes through the API because I’m integrating this into our own tools. The API has been straightforward, and I was able to get things working without opening support tickets.
The agentic actions are where I’m seeing a lot of the value. One question can trigger several steps and bring back a more complete answer. In cybersecurity, it’s critical that the response is grounded in our own guidance and shows where the information came from. CustomGPT.AI has been doing that in my testing. Review collected by and hosted on G2.com.
I’d like better visibility into how questions move through the multi-agent setup. When Ask Ryan routes a question across agents, I want to see which agent handled each part and how it contributed to the answer.
Right now, finding that information takes more digging than it should. It’s a specific improvement, but it’s the one I keep coming back to as I build this out. Review collected by and hosted on G2.com.