Guess which modules are missing from this process? OpenAI + Google Ads

:robot: Lee este artículo en español :es: en LinkedIn:
Comprender la intención de las palabras clave de los anuncios de Google con OpenAI

:mag: Mystery Automation Game!

Guess which modules are missing from this process? :man_detective:

:star2: Hello to all automation and challenge enthusiasts!
I will show you an automated process created in Make, and your goal will be to identify the modules and actions missing from the flow.

Are you ready to participate? Please leave me your comments!

:dart: Process: «Understanding the intent of search terms in Google Ads campaigns»
:sunglasses: Objective: «Detect unwanted terms to add them as negative words»

:hammer_and_wrench: Tools used:
Make: the automation platform.
Google Ads : Google’s advertising platform.
⊙ A Google spreadsheet.
OpenAI: using the GPT3 model

:wrench: Process description:
1) Every Monday at 7 AM, a Google spreadsheet is generated.
2) Then, the list of clients and their Google Ads accounts are searched.
3) For each client, the search terms of the previous week are searched.
4) Each search term is classified according to its intent (commercial, transactional, or informational).
5) This information is saved on a different sheet for each client in the generated form.

:mag: And there you have an overview of the process!

Now, I challenge you to tell me which modules or actions are missing from the image.
Can you identify the parts that are not represented?

I look forward to your feedback and participation!
Let the game begin.

This is the result of executing the procedure, 1 Google spreadsheet with several sheets; each sheet contains the list of search terms in a Google Ads account.


Ohh wow, thanks a lot for putting this together @Francisco_Fontes, how exciting! :clap:

Let’s see which curious mind will be the one to untangle this mystery! :man_detective: :crystal_ball:

It Crowd Popcorn GIF

This is really interesting and much harder than it looks.

I have the rough idea, but also I’m not sure exactly what the JSON and group modules are for. But I don’t use Sheets much.

Anyways this is the process as I understand it:

First you use the sheet name to look up the client’s account and then iterate through each of the keywords and save them to the client’s sheet in the B Column.

Then you send each keyword to chatGPT to classify its intent.

Then take the chatGPT keyword classification and then use the cellID (or some other identifier) of the keyword to determine where it should go (the row column is cellID-1).

You must use the iterator to go process each keyword.


I am too tired to answer, but nice solution and approach nonetheless :stuck_out_tongue:

I share the resolution of the game.