---
title: "Comment by Charlie from xSquads on \"90% of the AI Agents Clients Ask For Are Just a Cron Job and One API Call\" | Hashnode"
description: "Same pattern here. The one thing I'd add before calling it done is the boring failure handling, because that's where the 3am calls actually come from: Store the"
url: "https://hashnode.com/posts/90-of-the-ai-agents-clients-ask-for-are-just-a-cron-job-and-one-api-call/6adc547f5860f38338b0c2c5/comment/6adc75cced30a65ec0215920"
---

[Discussion](https://hashnode.com/posts/90-of-the-ai-agents-clients-ask-for-are-just-a-cron-job-and-one-api-call/6adc547f5860f38338b0c2c5)

Comment [90% of the AI Agents Clients Ask For Are Just a Cron Job and One API Call](https://hashnode.com/posts/90-of-the-ai-agents-clients-ask-for-are-just-a-cron-job-and-one-api-call/6adc547f5860f38338b0c2c5)

[Charlie from xSquads](https://hashnode.com/@charlie-xsq)

[Oct 24](https://hashnode.com/posts/90-of-the-ai-agents-clients-ask-for-are-just-a-cron-job-and-one-api-call/6adc547f5860f38338b0c2c5/comment/6adc75cced30a65ec0215920)

Same pattern here. The one thing I'd add before calling it done is the boring failure handling, because that's where the 3am calls actually come from:

1.  Store the last-seen review ID instead of "now minus 60 minutes". If cron skips a run or the box reboots, a time window silently drops reviews.
2.  Make the model reply in JSON (`{"label":"negative","summary":"..."}`) and treat anything that doesn't parse as negative. A malformed reply should go to a person, not get ignored by the `split("\n")`.
3.  Post a daily "ran 24 times, 31 reviews, 2 flagged" heartbeat to the same Slack channel. Then silence means broken, not quiet.
4.  Keep 20 real reviews as a test set and rerun them whenever you change the prompt or model.

Still a cron job and one API call, just one you can trust when nobody's watching. For me it only becomes an agent when the founder also wants a drafted reply.

[Sohail Khan](https://hashnode.com/@msohailkhandev)

[Nov 4](https://hashnode.com/posts/90-of-the-ai-agents-clients-ask-for-are-just-a-cron-job-and-one-api-call/6adc547f5860f38338b0c2c5/comment/6adc75cced30a65ec0215920#reply-6aeb7b2737bc09f408c9b3d0)

That observability point is the one most teams learn the hard way, because a cron job that fails quietly looks exactly like a cron job that has nothing to do. The detail I would add is alerting on a missed run, not just a failed one, since a job that never starts leaves no error behind at all. Do you treat a missed run as its own alert, or fold it into the same failure channel?
