Hello! Thanks for asking.

A comprehensive, step-by-step guide for this was published in our use-case repository, which should cover all your needs.

If you're interested in broader AI applications, you might also consider using Eden AI as an interface for your chosen AI solution. They offer a unified API for various AI tools and a single billing point, simplifying the comparison of outputs from different engines. There's also a callin.io app available!

Hello!

I'm encountering an issue with aggregating call transcripts using the current tutorial. I've followed the template and used the provided JSON, but it seems only the initial sentence of the call is being captured.

I've verified the call itself, which is 22 minutes long, so the issue isn't with the call duration.

Could you please advise if there's an error in the tutorial or my setup that's causing this? I'm attaching the blueprint.json file along with screenshots of the module settings for your review.

Any assistance would be greatly appreciated!

Screen Shot 2023-08-15 at 10.32.58 AM312×623 37.3 KB

Screen Shot 2023-08-15 at 10.32.13 AM469×611 24.5 KB

blueprint.json (25.8 KB)

I can’t see any obvious error, but I don’t have any Gong call records in our sandbox to test with.

If you run the scenario, you can then click on the “speech bubbles” to the top right of each module to see exactly what it returned. Take a look at the results of the callin.io API Call module, the Iterator, and the Text Aggregator.

If nothing jumps out to you, copy and paste screenshots from each of those and post here.

Thanks! Nothing seems out of the ordinary, so I've included some screenshots of the workflow.

It appears the Iterator is only capturing a single sentence from the "Make an API call" module.

What are your thoughts?

Screen Shot 2023-08-15 at 11.28.31 AM487×608 42.9 KB

Actually, when I examine the Call Transcripts Array within Bundle 1, I can see all the bundles from the transcript. However, when I proceed to the Iterator, my only option is to select sentences, which results in only one sentence being returned, not the entire array.

Do you think there's a workaround for this?

Screen Shot 2023-08-15 at 11.39.09 AM506×618 50.3 KB

Screen Shot 2023-08-15 at 11.40.47 AM745×538 63.2 KB

Yes, that appears to be the case.

For the Make an API Call module, please expand the callTranscripts element completely.

Within that, you should find a transcript array, which contains an array of sentences.

Yes, I understand.

However, when I select sentences for the iterator, as shown in the attachment, it's only returning one bundle in the array.

I suspect this is due to the transcript from the API call not reaching the iterator for some reason.

I'm not entirely sure if there's anything else you can do without perhaps scheduling a call to demonstrate.

Unless you have any other suggestions, would you be open to arranging a call?

Thank you very much.

Hi there.

It appears there might be an issue with the initial use case setup.

The Gong API documentation indicates (under Response, when expanding each "...") that the callTranscripts array contains one entry per call. Each entry includes a transcript array, with one entry per monologue, which in turn holds the sentences from that monologue.

The term "monologue" is significant here – a new transcript is generated each time the speaker changes. The example in the use case only captures sentences from the first monologue, which in your scenario is limited to "Can you hear me?".

To address this, you'll need to insert an additional Iterator module between the Make an API Call module and the existing Iterator module.

The new Iterator will be responsible for processing the transcript array:

The original Iterator should then process the sentences from the newly added Iterator:

image1070×403 33.9 KB

Finally, the Text Aggregator needs to be configured to aggregate all items from the new Iterator:

image633×883 49.4 KB

If this adjustment resolves the issue, please confirm, and we will proceed with updating the published use case.

IT WORKED!

Thank you so much for this - it's incredibly helpful for our team!

So glad we were able to get this working.

I appreciate all of your help!