AI role in recruitment of right-fit talent

Organizations are transforming themselves by leveraging AI tools and technologies to ensure they bring in right-fit talent into their firm. They are looking at optimising certain processes, making it effective and faster, making it cost effective, trying to automate steps to the extent possible, bring in more clarity and intuitiveness wherever needed and applicable and therefore the list goes on. Some use cases are described below which are reflections to these problems.

1. Defining appropriate content in a Job Description - This is not simple and all about so called "scope". It requires thinking aspect, what exactly one needs as part of the role and hence putting up a content that is not bookish, instead more pragmatic and needed to augment the role for "success". AI algorithms and mechanism can help learn from past historical data, trends, what works for firm and what does not, which skills are in demand and specifically to the success and roadmap of priorities of firm from a business strategy / mission standpoint. We have seen many examples where "must have skills", "nice to have skills" are often compromised and a laundry list of many elements are captured. At the same time, many times these are articulated in a refined manner. The key is to make it crisp, linked to business goal.
2. Screening of candidates - Some of the study / research indicates that more than 50% of talent acquisition teams indicate that screening of candidates are one of the most time consuming tasks and toughest considering variety of flavours of profiles available to them to choose from. Hence it becomes increasingly important to have a mechanism which can validate based on certain rules and more importantly use AI to see if this process can be learned from past data and AI can help in minimising ambiguity. There are of course ever growing challenges to this as well. Over past few years, automation of these searches, classifications are possible due to defined, clear rules. We need to leverage modern machine learning solutions to be able to perform these tasks where rules may not be clearer, harder to find such business rules etc.
3. Right-fit candidate pooling, identification and recommendation - This is another use case which could fall into the "screening" process. However leveraging AI to recommend these are important and critical considering ever changing role of requirements, changes in needs to manage from business priority perspective. Lot of parameters also need to be considered where gender, age, name and other biased parameters (if any) can be removed during data preparation step to ingest into a pool of shortlisted frame which can be fed into the modelling process to derive recommendations.
4. "Test" appropriateness on Content, Newness, Validity - While lot of materials and tests are prepared, created, maintained and used for initial screening purposes, it is important to make the content as per business priorities, aligned to business need, overall strategy of the firm, how the particular role will impact to the success of firm's vision, how is the content relevant, is it constantly updated to fit the "newness" aspect needed for job at hand and at ground level etc. AI algorithms can help refine these based on active learning and feedback loop process.
5. Video analytics / Gesture analytics - Typically the face to face discussions are not captured and analysed for future improvements. This is very critical, not only to get a clarity of assessment for the candidate, but also to improve one's own process of flow of discussion aligned to the "business need". Process can be defined and thought around using AI where behaviour interpretation of the candidate, conversation pattern, willingness to learn new technologies, flexibility to get into the culture of organisation can be explored and analysed. A simulated office environment can be presented to candidate to assess very practical way of interactions that happen during the process, actions / reactions, confidence level, emotional quotient etc can be looked at for more clarity. We all know how important these "soft aspects" are important in a "successful business environment" rather than focusing on only few specific aspects.
6. "Test" credibility during a web conference mode of testing - It is important to understand how candidate thinks, looks at a particular set of questions, confidence level, trustworthiness, eye movements / gestures / behaviour patterns while these are tested in a remote web mode most of the time. AI with the help gesture analytics can help recognise patterns, help detect anomalies and provide a score more from a trust and credibility point of view in addition to how one is performing in the test.
7. Assessment of quality of coding - Skills are critical and know-hows are very important and should be very specific to the need of the role. AI can help assess in a manner which will learn from past datasets, patterns and to be able to recommend better and more crisp assessment of steps in future. Secondly, lot of these steps can be skipped provided certain assessments, certifications are already done and proven for a role by the candidate in the past and AI algorithms can search during screening process to flag this aspect for quicker turnaround if need be.
8. Variations based on experience level, category of role, need from business priority perspective - Entire process will vary based on some of these factors or more parameters like these and could be a combination of some or all. Intent is to get more clarity based on relevance of these parameters. The variation aspect score can be an appropriate indicator for faster decision making.
9. Social impact by content / subject - There is a need to understand and analyse whether a particular candidate has displayed some sort of exposure in social media, in what sense and what is the impact generated from context. E.g. association with some of the social media sites would help trigger the interaction level, frequency of interaction and then more importantly what is the output of association. Some parameters to assess and analyse could be - blogs, point of views, recommendations, publications, papers, participation in hackathons, highlighting lessons learnt, responses to a problem and most of these related to which domain, what technology area, story telling approach etc. (and all in a quantitative and qualitative aspect that is relevant to business need of the firm and direction in which their roadmap is heading towards..)

While there are thoughts around how the AI / ML process works internally, is the learning process better and effective and so on, explainability of most of these will depend on inputs we provide to these algorithms, approaches / steps we define from our need, parameters we provide, rules we set, historical information we give, and hence variations in data will have impact on output coming from model based on statistical significance and other aspects. Decision making will have to be ours based on the information and ammunition we have from AI/ML output.

Organizations usually have to spend a lot around the entire recruitment process for the "right-fit" all the time and this is a never ending need. This will continue to exist and demand supply pattern will continue to get optimized. AI and ML algorithms will make things clearer, quicker and better in the process of overall decision making for people on the ground to take decisions.

Disclaimer: "The postings on this site are my personal point of views from my experiences, thoughts, readings from various sources and don't necessarily represent any firm's positions, strategies or opinions.”