Question: My workflow: Incoming Gong call transcript, AI agent runs and summaries and scores call based on set parameters. Then it sends out alerts and appends data to a sheet. There are 3 agent steps 1st agent does summarising and scoring 2nd agent does feature extraction i.e if a person mentions some specific keywords 3 agent finds moments of love or antilove i.e where we had a real impact on the customer vs where we dropped the ball I have set up 2 lindys doing exactly the same thing but their structure is different Lindy 1 > all the above agents are in 1 big workflow (ss1) Lindy 2 > All agents have distinct workflows but share the same trigger ie. incoming Gong transcript (ss2) My questions:
Which Lindy will be more credit efficient?
In SS1, see the red highlighted box, when the flow runs it should ideally move to the next step irrespective of whether a relevant product feature mention was found or not. But the flow stops there without going to the next step?
Marvin A. any thoughts?
hey Aayush M. — thanks for the detailed breakdown! Cool use case btw. on the credit efficiency side: it’s honestly really hard to predict that just from workflow structure. i’d recommend testing both with a few examples and comparing usage. also linking you to our guide on reducing credit usage — might be helpful! for the second issue: that sounds like a prompting thing. if the agent’s task is clear enough, the default exit condition (exit once the task is done) should work fine — that’s what we use in most flows. if needed, you can always add more granular exit conditions, but from what you’re describing, tweaking the prompt should probably do the trick.
lmk if thats helpful!
Um I tried this but the agent for some reason stops working 😕
