Why are we still setting up AI agents manually?
We talk a lot about AI automating work, but setting up AI products still requires a surprising amount of manual effort.
Take an AI customer support agent. You need to connect integrations, configure instructions, define business rules, set permissions, test different scenarios, and keep everything updated.
What if another AI agent could handle most of this?
You explain what your business needs. It asks questions, connects the right tools, configures everything, runs tests, identifies issues, and prepares the system for your approval.
And when something changes in your business, it can update the configuration and test it again.
I have been thinking about this approach while working on AI products. It could remove a lot of the complexity that makes powerful software difficult to adopt.
But I wonder whether people would actually trust an AI to configure another AI.
For those building SaaS products, would you give an AI this level of control over your product setup? What would you still want to configure yourself?
Replies
Part of it is that every agent tool has its own config format. I maintain a small plugin that installs the same skill into 8 different coding CLIs, and most of the work was mapping one setup to eight folder layouts. The instructions were the easy part. A shared standard would remove a lot of this.