Hey team,
Is it possible with how the agents are setup to have have one trigger from a slack message or something (notice there aren’t any slack triggers, just actions). And then have it feedback messages in slack like an agent does on other platforms. And then when it has a question it holds and waits for an answer (ideally temporarily spinning down while it waits so it’s not burning through credits if I don’t get back to it right away). I’m pretty sure I could do something like that with Relevance. But not sure if the latenode agent would be able to function in that way?
Hi Dan!
You’re most likely referring to a Human in the Loop type of mechanic - it’s not natively supported in LateNode at the moment.
It can be set up, but only through a more complex flow that involves saving the intermediate state of the scenario (not the agent) in a database.
In this setup, the system records when the agent has responded and is waiting for user confirmation; then, a separate trigger can continue the scenario in a subsequent run, within the same session and using the same request ID.
By the way, Human in the Loop is already on our roadmap, so this functionality will definitely be added to the platform.
I can’t share exact timelines yet, but it’s planned for future releases.
Also, could you tell me a bit more about your use case? I’m really interested to hear what exactly you’re trying to build.
And by the way - Slack triggers do exist, they’re just located in second folder.
At the moment, there are two versions of the Slack integration on the platform. They’ll be updated and merged into a single version soon, but for now, you can find them in the “slack” folder, not in “slackbot.”
So basically my EA is resigning in December, and I’m weighing up if I get a new human EA or if I just try and build out a bunch of agents to replace the majority of work she does. So was hoping to have it on slack to start certain tasks… It might be like - here is an invoice, can you load it into Xero. Which probably it could do most of those sorts of things without much human middle stuff, but I was hoping that it’d ask me a question if it was stuck on something or had some edge cases. Then was thinking I’d have a bunch of sub agents it could pull in under the master “EA” agent kinda thing.
You can definitely work with this. If you have any questions about the logic or implementation, please feel free to reach out to us, and we will do our best to explain everything to you.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.