I recently discovered a highly successful automation I created a few months ago that has been generating over $15,000 to $20,000 every month like clockwork. This system consists of two main components: gp4 open AI and air table. The former filters and processes jobs for me, while the latter stores all the records.
The automation begins with an RSS feed module that connects to Upwork's RSS feed feature. This allows me to create specific job queries using Boolean connectors and query the RSS feed every hour or two for matching jobs. I set up the URL with a query to return 200 items at a time, sorted by recency.
The RSS feed then goes through a parse description HTML to text module, which turns the HTML description into plain text. The next module is the secret sauce of the entire system: a filter job module using gp4 open AI.
The filter job module prompts the AI to filter jobs for relevance, write a short introductory icebreaker, and note API integrations. I provide examples of input and desired output for the AI to learn from, making it perform better.
After the filter job module, I check to see if the job is relevant. If it is, I get the budget of the job and set the payment model to either hourly or fixed price. I then add the job details to air table, including the title, description, date posted, introduction, and a template for the job application.
The benefits of this automation include tracking job proposals and money made from them, using flexible AI filtering, proceduralizing the icebreaker process, and providing the ability to record a Loom video for high-value jobs.
This system has been a game-changer for me, allowing me to maintain a high volume of job applications while still maintaining a reasonable level of quality and providing the ability to record Loom videos for high-value contracts.