Dumping a continuous firehose of Slack messages into an LLM context is a guaranteed way to bankrupt the company. Plus, the latency is a joke: by the time the agent parses the thread and hits the Jira API, you could've manually dragged the tickets across the board three times over yourself
I don't need lower latency. I need to not have to do it myself.
For that to work, though, I also need to not have to check it myself. The tool has to be reliable.
That's exactly the trap with any LLM agent in prod. To completely drop the human-in-the-loop, you need a confidence level that current transformers simply lack architecturally. The second the model silently closes a release ticket because someone made a sarcastic joke about a rollback in the thread, you'll be the one rushing to rip that tool out of your stack
This is literally an LLM comment, the structure is the same.