RPA+AI utilizes various AIs for automating pipeline processes
Robotic Process Automation with make.com
Intro
With make.com, we can create a codeless AI pipeline, allowing the outputs of different upstream models to seamlessly integrate into downstream models. A simple example would be, I share a LinkedIn post, where I briefly summarize the content of the latest article I’ve read:
Method
This workflow can be used to send updates either instantly or at scheduled times. In this example, I have set up an xls file with two columns of data. The first column consists of a series of URLs for articles or videos that I have already viewed. The second column is a simple tagging column, where all links tagged as “latest” will be processed by the Perplexity API (note that you must be a paid user of Perplexity to use the API).
The processed results will be input into OpenAI GPT for editing LinkedIn posts. The final step is to publish the post on LinkedIn.
The whole process is linear, and then make.com offers various types of Flow Control, such as Router, which can be used to handle multiple terminals at the same time. For example, we not only want to post to LinkedIn, but also send the Perplexity output to the Anthropic model, and create a Facebook post.
For further information, please refer to this video.