Alper Tayfur

I m not talking about simple reminders or moving one row between apps.

I mean the messy workflows that involve several steps: checking information, making a decision, updating another tool, asking for approval, sending a message, and making sure nothing was missed.

Have you ever tried automating one of these workflows with AI agents or traditional automation tools?

If yes, where did it become difficult: connecting the apps, keeping context, handling exceptions, avoiding duplicate actions, or knowing when the agent should stop and ask you?

Went through today's launches properly this morning. Generate this, automate that, no code needed. Every page is a capability list. Not one of them tells you what the product does on the day it gets something wrong.

That's backwards for anything a person puts their name on. The failure that actually costs you isn't the crash, because a crash announces itself. It's the confident wrong output that ships looking fine. A video that publishes a false claim to someone's channel. A generated model that's subtly the wrong shape. An agent filling in its own verification field.

Building an AI agent is one challenge, but knowing it's ready for production is a completely different one.

Traditional software can often be verified with unit tests and integration tests, but AI agents introduce additional complexity. They rely on reasoning, external tools, changing context, and non-deterministic model outputs, which makes testing much less straightforward.

I'm curious how other teams approach this before deploying AI agents to real users.