Skip to main content

Hi, I’m a content marketeer at a SaaS company. I love Gong. I find myself reading through the call transcripts of our sales and csm people quite often in search of interesting topics, returning questions, pain points and detractors of prospects and customers. While I don’t mind doing this, it is time-consuming and difficult to prioritise or segment ideas, because I can only review them one at a time. I’m wondering if there is a way to automate the process and doing it in bulk to generate a better overview. I have been looking into creating a flow using Zapier, ChatGPT and Notion, but maybe there’s a more straightforward approach available within Gong itself. 

Hi ​@Geert Merckaert! I believe that you can accomplish this with the Gong Data Cloud, Snowflake, and other Gong integrations. ​@Andrew O'Driscoll do you have ideas around this?

Is there any way to export in bulk transcripts into google sheets via api?

Same as Geert, i’d like to do a meta analysis on select gong call transcripts. Ideally i would do that in Gong but maybe there are other tools better suited for that?

Curious if there was an outcome of this? Would like to see trends at a higher level as well. 

 

Hi there ​@Olivier Decroupette ​@Lauren Troise — the AI Theme Spotter agent can do this for you. 

I hope this helps!

​@Geert Merckaert our marketing department approached us with a similar request. We are using Hubspot which by now has a native ChatGPT integration which can promt every property, meeting or call summary content for example which is being synced from Gong to Hubspot.

If we combine that with a specific smart tracker looking for frequently asked question for focus on those meetings the outcome should be useful, as we can output the answers into a table outside of Hubspot and then let a bot run over everything to identify patterns.

It would be great to do that on transcripts, but thats not really possible yet. Will see what the AI theme agents in Q4 do, but for now I am exited to see how far we can get with it.

Maybe it is something you could try?

I’m doing this with Clay and Humata right now. Use the native integration Gong calls to a table (you can also do this one rep at a time). I haven’t figured out how to limit the Clay inputs to only specific streams, yet.

Then I do some filtering down on all those transcripts so I’ve got a few Clay tables fed from the main table. Right now I’m doing the separation by the main purpose/product discussed on the call.

Then I extract the transcript and have some prompt columns where an LLM extracts questions asked, objections raised, interesting ideas or strong opinions.

I found that was still a little difficult to analyze in aggregate, but better than flicking through one gong call at a time.

So then I turned all the transcripts to .txt files and imported them to Humata (runs automatically once a week). Humata is kinda like NotebookLM - lets you query large amounts of data in aggregate for themes. So I use that to identify most common objections, feature requests, questions asked, etc.

Reply


Terms & ConditionsCookie settings