Building AI agents with a visual programming environment

There is a visual programming environment that allows users to compose complex programs like AI agents without the need to interface directly with code. It's not no-code, just a visual layer on top of code for better composability and observability. At runtime, every event triggered lights up on the canvas, appearing like neurons firing in the brain of the agent. This isn't just there to look cool - it allows easy debugging without having to read through hundreds of thousands of lines of logs and code to diagnose and fix problems. Building agents is also simpler because it ships with a kit of parts that you can use to assemble an agent, using the embedded agent. Interestingly, that embedded agent itself is built with the same framework and can be opened and edited (pretty meta - the agent can be used to make improvements to itself), allowing full control and customizability along with a template/reference. Interested to hear your thoughts on this - do you think this approach would help you trust an AI agent more? Disclaimer: I am part of the team working on this product, but I'm genuinely curious to know what this community thinks about this approach to building agents.