[Student IDEAS] by Yinzhe HUANG - Master in Data Sciences & Business Analytics at ESSEC Business School and CentraleSupélec

Digitalization and AI-based applications are being integrated into firms’ human resource management (HRM) approaches for managing people in domestic and international organisations. Amongst the whole process of HRM, Talent Acquisition (TA) is the best-equipped one with Digitalization. There is no doubt that machines are good at some sort of work. However, we must know that the efficiency sometimes takes costs, such as privacy and bias.

Talent Acquisition Process Review

In general, TA can be splitted into three main steps: planning, sourcing, and selection.

• Planning: a process in which an organisation attempts to estimate the demand for labour and evaluate the size, nature and sources of the supply which will be required to meet that demand [1], including knowledge, skills, abilities, and other characteristics [2].
• Sourcing: the use of one or more strategies to attract or identify candidates to fill job vacancies, including posting job requirements on media, using headhunters, and employee referral.
• Selection: assessment of person-job fit, i.e., the competency power of candidates on particular job requirement, usually consisting of a series of funnel-like filtering processes including resume reviewing, testing, interviewing and due diligence.

Meanwhile, the processes of TA are not separated. In practice, the processes work in a PDCA way, with continuous iteration and evolution as the business moving forward.

Digitalization in TA

To visualise the comprehensive picture of how digitalization technologies affect TA and human-machine configurations at work and their influences on organisational level outcomes, we develop an overarching conceptual framework through this review:

In general, the PDCA circuit [3] is essential in the organical combination of Talent Acquisition & Artificial Intelligence. AI assists in HR planning by determining future employee needs and making effective recruitment decisions [4]. It is also evident that AI-enabled recruitment and selection play a crucial role in attracting and selecting the most talented work pool to the organisations, as these advanced technologies can access data and make decisions at a speedy pace and can handle large volumes of information in a time that far exceeds human capacity [5]. As a result, AI algorithms can improve job candidate identification, that is who is most interested and suited for the job and provide better communication of the job opening. Influencing job seekers’ technology increases their participation in AI enabled recruiting [6]. AI also assists in making the job interview process more effective, so that the interview process is now changed from face-to-face to internet-based interviews, such as asynchronous video interviews (AVIs) . Moreover, Pessach et  al. [7] found in their studies that using a hybrid decision-support tool helped HR professionals in the recruitment and placement processes and increased the impact of recruiters and maximised organisational return on investment. AI Algorithms allowed HR professionals to identify suitable profiles for job vacancies, eliminating cognitive biases of race, gender and sexual orientation that mar human judgement in recruiting activities. The latest work of Baidu [8] even takes the salary management into account.

In the following paragraphs, we will take two examples from the Chinese context to show how system design and AI can empower talent acquisition.

Case study 1: Lark as User-friendly TA SaaS (Powered by ByteDance)

Before talking about how AI upgrades automation in the talent acquisition process, we would like to briefly introduce the key elements of a user-friendly digital talent acquisition system. In this section, we would like to take Lark (an enterprise cooperation platform developed by ByteDance) as an example. 

• Role Definition & Permission Management: In the talent acquisition process, the stakeholders include not only Human Resources Managers, but also many other roles, such as supervisors, administrators, and headhunters. Therefore, it is essential to assign different levels of permissions to different users and roles. In Lark TA, different interfaces are prepared for different roles. For the supervisors of a particular position, what they need to know is just the information related to their subordinate position. Meanwhile, we also need to separate external users of the system, such as headhunters, from unnecessary internal-used information. 
• Cross-platform Cooperation: Although an All-in-one platform is too ambitious, a tool would be much more user-friendly if it is good at cross platform cooperation by preventing HR Managers from being occupied with dealing with transferring from systems.  An important role of the platform is a Talent Acquisition Pipeline. The pipeline starts from “data collection”. The data inputs of the systems are the resumes collected from different sources. No matter what kinds of formats of the resumes (pictures, webpages, text or documents), they should be present in the same way for anyone who is going to review them. In Lark, HRs are able to present a job description to most of the popular job-seeking platforms and assign budgets for advertisement. When the resumes are collected, both the HRs and supervisors are able to review the resumes with both the initial texts and the highlighted information. Afterwards, the platform should be well-coordinated with calendar management tools so that the arrangement of interviews can be more efficient. Normally, a company has already got its preference in the calendar management systems. The way Lark tried to get it well-organized is to develop add-in tools for cross-platform cooperation.What is more, Lark takes its advantages in voice-text translation system, it can automatically provide both the raw text and the summarizations after interviews so that different interviewers can got the context of previous interviews. Finally, the system provides automatic templates and an email-sending system to give feedback to the candidates.
• Talent Pool: Beyond a well-defined pipeline, the system should also accumulate a talent pool for talent mining. Even though candidates are not perfectly fit for particular positions, they may show their light in other job positions. The talent pool is the key of building a bi-direction connection between the corporation and the candidates. Beyond simply accumulating data, Lark also provided label filtering and a CV searching engine for the HRs so that they can pinpoint the candidates that are highly qualified for new job positions.

Case Study 2: Model-driven TA by Baidu Talent Intelligence Centre (TIC)

Founded in 2015, Baidu TIC aims to form a systematic data-driven solution for talent management based on the accumulated advantages in AI and Big Data of Baidu (HBR China, 2018). After 6-year cultivation in this area, the team developed many algorithms with real-world applications in Talent Acquisition.

Technical Backbone Overview

In terms of technical accumulation, the team started from the analysis of trend of job market [9] by dynamically tracking the evolving recruitment topics with hierarchical dirichlet processes. And then, the team evaluate both the profile of employers [10] and the skills of employees [11]. With those key insights being clarified, they built a Person-Job Fit Neural Network [12] Design, Operation, Product, and Technique) and enhanced the model by designing an Ability-Aware Mechanism [13]. Apart from the resume filtering which gets HR into loads of repeat work, the team also focused on AI-driven Interview Assessment [14], which enables a variety of applications in job interviews, and AI-driven Salary Management [15], which assesses skill value from a market-oriented perspective.

AI Penetration in TA Processes

Nonetheless, we shall always notice that even state-of-the-art techniques are the means to realize the end. It is more important to recognize the value creation than to simply show off IT skills. Therefore, the systematic design, which combines the value creation chain and the state-of-the-art techniques organically, is even more essential for business success.

With the technical backbone, Baidu is able to build a whole solution for AI-driven Talent Acquisition (CSDN) which can boost HRM extraordinarily. The solution concentrates on three main perspectives in HRM: talent, organization, and culture. In this article, which focus mainly on talent acquisition, we would like to emphasise more on the first part.

In the talent perspective, the solution covers almost the whole process of talent acquisition, including job analysis, sourcing, and selection.

• Job Analysis: Baidu TIC provided bottom-down support for the job analysis step. Based on the giant amount of internal and external data, TIC utilized an enhanced topic model on the trend analysis on the job market, supporting the strategic design of future talent acquisition. For example, TIC succeeded in predicting the mass demand on talents in Big Data and Artificial Intelligence would overwhelm that in Online to Offline, by modelling on millions of job descriptions. Meanwhile, it forecasted the strategic planning of competitors on Autonomous Driving accurately. With such support, Baidu became one of the companies that moved to road test of driverless vehicles. The job analysis provided key information for the strategic level, it also made the operational level more convenient. Thanks to the bipartitional (job description and resume) analysis, it was also able to extract key skills for different positions, so that HR can articulate the specific requirement more clearly. The job descriptions are better-composed.
• Sourcing: It is more on pipeline issues than on modelling issues in terms of sourcing. In the old ages, the resumes collected from all channels should be validated by Human Resource Managers. However, if we expect the information can be assessed by model, the well-built pipeline is an essential prerequisite, in which the plain text in the resumes can be cleaned, aggregated, and formalized. During the data engineering process, several key elements form the instruments of data manipulation. The first element is resume resolution, during which the text in the resumes is recognized (especially for the pictures of resumes). The second one is information recognition, aggregating the information by analyzing the contents and positions of text, so that the text can be corrected and grouped. For example, the information of professional experience and that of education background should be separated. The third one is formalization. In this step, the information will be placed in the right location in the table. After these periods, the plain text is transformed to well-organized tabular data which can be easily analyzed and evaluated. The giant talent database ensures further analysis and model construction.
• Selection: In the selection part, TIC focused more on the person-job fit, in which they devoted most of their efforts in the evolution of model. In the scenario of Internet Companies, TIC made a model that is fitful for 4 main division in the Internet Companies: Technology (programmers, data scientists etc.), Product (product managers), Design (UI designers, artists etc.) and Operation (community, marketing etc.). These 4 kind of job positions cover more than 85% of the headcounts in the Internet Companies. The person-job fit model take both the resume and the job description as the input. With end-to-end model design and ability-aware mechanism, it not only scores the person-job fit, but also bring the highlighted ability recognition.

Covering almost the whole process of Talent Acquisition, this solution has become a module of Baidu AI Services. In application, Baidu co-operated with Beisen to help CP Group to realize Digital Transformation in Talent Acqusition.

Concerns of Digitalization in TA

We evaluated above some digital evolutions in Talent Acquisition. However, it is important to realize that efficiency comes at some costs. In this part, we would like to evaluate some of the concerns of digitalization in talent acquisition. 

• If bias exists in previous hires, it can creep in the system and make it even worse [16]. It is because the model was trained on historical data in which the successful application would be considered the positive samples. Sadly, we can not remove the bias from neither the training process nor the historical. Relying too much on algorithm resume filtering will result in the extension and even the expansion of bias. In practice, companies with such algorithms usually add manual checks before saying no to a candidate, while this would reduce the operational efficiency, contrary to the initial objective of liberating the HRs from reviewing the candidate over and over again. Therefore, it is essential to change the status of zero-sum tradeoff. Utilizing a more explainable AI in the recruitment process may be a solution. For example, instead of giving direct outcome of person-job fit scores, using an intermediate model such as Named Entity Recognition to first evaluate the talent pool of candidates may help HRs to better make effective decisions efficiently.
• The system did lighten the workload of HR, but human-machine co-operation requires Business Process Reconstruction. No matter how state-of-the-art the algorithms are, they are still means to bring add-in value for the business. Above the algorithm, a well-defined and user-friendly system should be constructed. Lark provided a digital solution covering the whole process of recruitment, in which the roles are well-defined with different authorities, multiple sources are connected, templates for JD and emails are prepared, and data visualizations are up-to-date. Beyond machine learning algorithms, there are loads of work to do in order to get the recruitment process fully digitalized.
• The system should be built under the constraints of privacy regulations. The examples illustrated in the articles are China-based companies. As we know, in Europe, companies are under more strict regulation in terms of privacy issues. For example, under GDPR (General Data Protection Regulation), the corporations are not allowed to store personal data as long as the subject refuses to give authorization and the max duration of data storage is no longer than two years. 

Conclusions

Digitalization can improve the efficiency of talent acquisition  in two main aspects: well-defined enterprise cooperation systems and machine decision. The former way improves the workflow by integrating the whole process, extracting process knowledge for stakeholders decisions under context, and seamless cooperation with other tools. The latter way protects the HRs from manually evaluating the candidates one by one. 

However, several concerns arise under the digitalization. We have to make an ethical trade-off between efficiency and personal rights. The way to digitise the Talent Acquisition processes is to improve the efficiency in systematic design under the prerequisites of compliance to regulation and respect of privacy. Not only do we need the improvement of efficiency, but also should we cage the power so that it can remain under control.

References

[1] Reilly, Peter. Human Resource Planning: An Introduction. Report 312. BEBC Distribution, 15 Albion Close, Parkstone, Poole BH12 3LL, United Kingdom., 1996.

[2] Peterson, Norman G., et al. An occupational information system for the 21st century: The development of O* NET. American Psychological Association, 1999.

[3] Koiesar, Peter J. "What Deming told the Japanese in 1950." Quality Management Journal 2.1 (1994): 9-24.

[4] Karatop, Buket, Cemalettin Kubat, and Özer Uygun. "Talent management in manufacturing system using fuzzy logic approach." Computers & Industrial Engineering 86 (2015): 127-136.

[5] Torres, Edwin N., and Cynthia Mejia. "Asynchronous video interviews in the hospitality industry: Considerations for virtual employee selection." International Journal of Hospitality Management 61 (2017): 4-13.

[6] van Esch, Patrick, et al. "Al-enabled biometrics in recruiting: Insights from marketers for managers." Australasian Marketing Journal 29.3 (2021): 225-234.

[7] Pessach, Dana, et al. "Employees recruitment: A prescriptive analytics approach via machine learning and mathematical programming." Decision Support Systems 134 (2020): 113290.

[8] Sun, Ying, et al. "Market-oriented job skill valuation with cooperative composition neural network." Nature communications 12.1 (2021): 1-12.

[9] Zhu, Chen, et al. "Recruitment market trend analysis with sequential latent variable models." Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining. 2016.

[10] Lin, Hao, et al. "Collaborative company profiling: Insights from an employee's perspective." Thirty-First AAAI Conference on Artificial Intelligence. 2017.

[11] Xu, Tong, et al. "Measuring the popularity of job skills in recruitment market: A multi-criteria approach." Proceedings of the AAAI Conference on Artificial Intelligence. Vol. 32. No. 1. 2018.

[12] Zhu, Chen, et al. "Person-job fit: Adapting the right talent for the right job with joint representation learning." ACM Transactions on Management Information Systems (TMIS) 9.3 (2018): 1-17.

[13] Qin, Chuan, et al. "Enhancing person-job fit for talent recruitment: An ability-aware neural network approach." The 41st international ACM SIGIR conference on research & development in information retrieval. 2018.

[14] Shen, Dazhong, et al. "A joint learning approach to intelligent job interview assessment." IJCAI. Vol. 18. 2018.

[15] Sun, Ying, et al. "Market-oriented job skill valuation with cooperative composition neural network." Nature communications 12.1 (2021): 1-12.

[16] Winick, Erin, "Baidu is testing neural networks that can match job seekers to jobs". MIT Technology Review (2018)