What's the last thing your agent still hands back to you?
We build agents all day, and the thing I keep noticing is that they almost never fail in the middle. They fail at the very last step, when something has to happen in the real world.
Ours kept dying in the same spot: a signup form that texts a six-digit code. Everything up to that point ran fine, and then it just sat there waiting for a human to go and read a text message.
I'm curious what everyone else's version of this is. Not the thing your agent does badly - the thing it does perfectly, right up until it has to stop and hand it back to you.
Verification codes? A booking that only happens by phone? A supplier who answers email once a week? Something that needs an actual human voice on the other end?
Mostly I want to know whether the wall sits in the same place for everyone, or whether it moves depending on what you're building.
Replies
its usually some tiny real world step for me like answering a call or confirming a code. funny how that ends up being the hardest part.
i have had this happen with account verification more than once. it makes you realize how much of the real world still sits outside the APIs can access.
I've noticed this with payments too. Most of the process can be handle automatically but the final approval still puls you back in,
the interesting part is that these arent really agent failures. the workflow is working fine it just reaches a step that still needs human access or judgment.
I would say phone calls are an even bigger wall for me. an agent can prepare everything beforehand but having an actual conversation still puts me back in control.
Account verification is definitely one of those walls. The whole flow can be automated until a human has to prove thy're actually there,
I have noticed the same thing with agents. They can handle a surprisingly complex workflow, but one simple werification step can bring everything to a stop.
Ours stalls on CAPTCHA every time. Everything else runs clean until that one checkbox needs a human.
What I find interesting is how close the agent gets before handing things back. sometimes its 95% done and the last 5% is the only part i actually have to touch.