top of page
Profet AI Insight

Know about the latest news and industry views through us

What sparks can AI and HR create?


"Recruitment" is the hottest HR topic in recent years. Facing market changes, competitions from international giants, generational gaps in the workforce, and various other challenges, a company not only has a significant increase in employee turnover, it has also seen more difficulty in finding suitable talent, let alone retaining them. Employee turnover not only directly leads to additional recruitment and training costs but also results in the loss of accumulated experience and knowledge within the company.

According to the Society for Human Resource Management (SHRM) in the United States, the cost of recruiting a new employee is approximately two-thirds of the total annual salary of the departed employee. Furthermore, a research report by the Center for American Progress (CAP) goes even further to suggest that for senior or highly skilled employees, the hidden costs of their departure can reach up to 213% of their total annual salary.

Fortunately, with the increasing maturity of digital transformation in businesses, rapid advancements in AI technology, and the rise of "AutoML" (Automated Machine Learning), we now have new solutions for the challenges in this new era. In this article, we will discuss how Profet AI utilizes AutoML to assist HR in making decisions regarding recruitment and talent retention. Additionally, we will further provide two specific case studies for readers to reference. These case studies aim to provide practical examples of the application of AutoML in HR-related processes.

The Application and Benefits of AutoML in HR

Regarding the application and benefits of AutoML in HR, here is a simple summary and description:

  1. Recruitment and talent retention: by predicting retention rates, length of tenure, or assessing the fit between employees and specific positions, as well as analyzing the turnover rates of senior employees, companies can gain insights into which talents are suitable for retention and whether they are well-suited for their roles. This information can assist HR in implementing innovative talent strategies and workforce planning.

  2. Training aspect: AutoML can be utilized to predict key talents and successors, as well as perform causal analysis. For example, through the use of the models, HR can identify the key factors and abilities that contribute to someone becoming a critical talent or a potential successor. HR can then provide targeted training programs for employees who possess these abilities and demonstrate high performance and retention rates. This approach enables innovative training for these desirable skills.

  3. Data-driven decision-making: In the past, HR decisions were primarily based on experience. However, AutoML can assist HR by providing data-driven decision recommendations. It can also delve deeper into the key factors influencing decisions and continuously optimize the decision-making process.

  4. Sensitive Information: HR data is often confidential and sensitive. In the past, it was challenging to outsource HR information to other vendors for modeling purposes due to the lack of suitable tools. However, AutoML allows HR to regain control over problem-solving by enabling them to leverage the power of modeling while keeping sensitive information in-house.

Case One: "Retention Rate Analysis" for Recruiting New Employees

When it comes to recruiting frontline employees in a factory, one of the most common challenges that HR faces is the high turnover rate. Despite going through the process of posting job vacancies, reviewing applications, conducting a series of interviews and training, new employees often leave shortly after starting their jobs. Following that, the same process needs to be repeated all over again. If HR could know in advance the expected retention rate or length of tenure for each suitable candidate during the recruitment decision-making process, it would significantly reduce the additional costs caused by the rapid turnover of new employees. With Profet AI's virtual data scientist platform, HR can obtain the desired results by simply preparing basic data and forms, and clicking a few times. Here, we provide a simple case study based on the problem-solving process of AutoML:

1.Problem Definition: before utilizing AutoML for modeling and problem-solving, the most important task is to define the problem that needs to be solved. Here, we want to solve

a.Problem: "Improving Employee Retention Rate"

b.Solution: "By predicting the retention rate, recruit applicants with higher potential for retention."

2.Preparing Relevant Data: Profet AI is an automated modeling platform designed for "structured data." It leverages historical data available to HR, such as basic background information of recruited employees (e.g., age, gender, commuting distance), details about the applied positions (e.g., plant location, department, process), as well as data related to whether the employee "remained after the probationary period" and "length of retention until departure." Simply organizing the data into a simple format allows you to begin automated modeling and analysis using the tools provided by AutoML.

Example dataset for predicting retention rate (columns can be added or removed based on the available data; here, we provide a simple example of the data format):


3.AutoML Automated Modeling: With the Profet AI platform, you can obtain modeling results comparable to those modeled by a data scientist without a background in statistics or data science. You just need to click on your mouse.

4.Result Application: The most crucial aspect of any AI or model lies in "how it is applied." Once the modeling is completed, you can fill in the information from the aforementioned table for the applicants you are considering to hire. Afterward, the system will automatically provide a "retention rate" result. In the case provided below, it indicates that there is an 80% probability that the employee will continue to be retained after the probation period!


Case Two: How to Use AI for Analyzing Employee Attrition Rate

Another challenge that HR faces is the attrition of senior employees and high-skilled technical staff. The development of key employees often requires years of investment, but these employees are likely to be poached by larger companies once their skills mature. This often results in smaller companies becoming training grounds for larger enterprises. With AutoML, is it possible to build a model that enables HR to identify which employees are more likely to leave or be poached in order to propose appropriate countermeasures in a timely manner? To address this problem, we repeat the steps mentioned above:

1.Define the problem:

a.Problem: "Reduce Employee Turnover Rate"

b.Solution: "By predicting employees with a higher likelihood of attrition, proactively propose countermeasures."

2.Prepare Data: gather relevant data. Here is an example:


3.AutoML Automated Modeling: With the Profet AI platform, you can obtain modeling results comparable to those modeled by a data scientist without a background in statistics or data science. You just need to click on your mouse.

4.Application of results: After the modeling is completed, the information in the above table can be regularly filled in for existing employees. The system will give a percentage of the possibility of employee attrition. The usual method is that HR regularly updates the above information to obtain information related to the attrition rate, and then conducts additional talks or uses other strategies to retain employees with high attrition rates.


As the application of the model matures, new features will continue to be added in the future. In terms of "salary," not only the absolute value of salary has an impact, the salary of the same job in the market is also an important reference indicator. If you are not sure how to implement additional features, Profet AI also provides after-sales consulting services, which can be used according to the company's needs.

Although I understand the above cases, I am still not sure how to start. What should I do?

If it is still difficult for the company at this stage to directly achieve the above content, there is no need to worry. Profet AI also provides a workshop cooperation mode such as "AI Hackathon." In this solution, Profet AI's professional consultants will provide a course on AI general education, topic search, and hands-on modeling, so company members can quickly understand what AI is, and what application cases are there. It is more important to cultivate the concept of "data application" and learn how to apply predictive models, therefore the culture of AI and data-driven decision-making can take root within the company, and ultimately help reap the fruits of digital transformation.



bottom of page