Use these resources to learn about query creation in DevOps

ChatGPT at Work: A Cheat Code for Productivity

Program Granite,Innovative Solutions,Productivity

July 11, 11:30 AM July 11, 01:55 PM

In this article, I am going to discuss how you can use OpenAIs ChatGPT API to automate monotonous tasks and skyrocket productivity, saving countless dollars and labor hours in the process.

Earlier this summer, I was searching for things to do at work and was given the task to help the HR student interns make job profile summaries using existing job descriptions. At first, I was delighted to help but after finding out that there were more than 1500 summaries to be created, I thought that there must be a better way. To paint a picture in your head of this monumental task, I will describe the process of creating one summary. You first had to look at a spreadsheet, read all the essential functions of a job, and then synthesize the essential functions to be under 100 words. After reviewing some summaries that were already finished, I found that the interns were not spending enough time reading, internalizing, and producing meaningful summaries that were descriptive enough to be featured on an official job listing. But was it their fault?

If you were asked to read 1500 paragraphs, condense them, review them, edit them, and then send them to someone else to revise, all while under the pressure that these would be seen by potential BYU employees, you would simply get burned out. Quick.

So, because of my ingenuity (a fancy word for laziness) I copied the essential functions of a job and put them into chatGPT, asking it to summarize the text into a 100-word paragraph. To my surprise, they were not bad, and they were far more comprehensive than the man-made summaries. After showing them to the functional lead in charge of this project, she said that AI generated summaries great and that we should transition our efforts into automating all 1500. After a day of sporadic programming, I created a python script that iterates through a list of job grids, converts them to CSVs, and feeds them to the ChatGPT API, placing all the finished summaries in a Word document. After conducting some tests on the efficiency of my script, we discovered that to do all 1500 job profile summaries, it would take 29 minutes, and would cost roughly $3.50 in API call expenses. Considering that Program Granite was willing to pay two interns $15 an hour over the course of four months to complete this project, I would say that we have saved a great deal of time and effort.

The biggest roadblock we encountered amid our exploration of the chatGPT API was the potential risk of sharing proprietary information with Open AI. BYU does not have any Data Sharing Agreement (DSA) with Open AI so anything that we feed the ChatGPT web interface, or the API gets stored in their servers and is used to train their learning models. When the Program Granite leadership committee heard about our little project, they asked that we place it on hold until we know more about the risks of sharing potentially sensitive information with a third party. After talking to some of BYUs head IT people, they determined that if we cleaned the Job grid spreadsheets and got rid of the private information, they would not be considered proprietary since Job descriptions are already available to the public. This was a massive breakthrough as we could now continue our project to summarize all 1500+ job descriptions and eventually implement them into Workday.

As an information systems major, I actively search for opportunities to combine the fields of computer science and business to make better decisions, optimize my work, and save resources. As the world embraces AI and machine learning, it is crucial to understand how they work so that we can continue to create value for clients and improve the efficiency of our business processes. The example that I shared in this article is a small but effective way of how new technologies can be used to boost productivity and save money. I hope that anyone who reads this article will seek out and find opportunities to improve their technical skills and use them to benefit their teams.