The expectations and hype that we see today around artificial intelligence (AI) is simply amazingly huge. Soon we will be talking to our computers, drones will do our shopping for us, cars will start to drive themselves, and most office workers will only control the operation of the machines. Is this true and how realistic is all this?AI RECRUITMENT SOFTWARE.

As an industry analyst and engineer who has been studying technology for decades, I can say that we are going through a rather interesting stage when, on the one hand, the hype around is far ahead of reality, and on the other hand, the result may turn out to be much more significant than we think. Well, the opportunities at the level of personnel management are simply huge.

Even though almost all HR providers are working on building AI teams and we all want our system to become smarter and more efficient, I feel like the current market is still too young, and to prove it, I would like to highlight a few points.

Not too long ago, I attended a recruiting automation conference where Billy Bean, CEO of the Oakland A's professional baseball club, spoke about the Moneyball book. After giving an amazing account of the history of sabermetrics and the impact data has had on baseball, he said that he currently has six doctoral machine learning engineers working for him, and “it’s pretty hard to beat a team of PhDs.” This is exactly what we are seeing in business today.

The role of AI in HR and leadership

Admittedly, AI is not some magical computerized personality, but a wide array of machine learning algorithms and tools that can quickly acquire data, identify patterns, and optimize or predict trends. The systems can recognize speech, analyze photographs, and use pattern-matching techniques to determine mood, honesty, and even personality traits. Such algorithms do not rely on "intuition" like a person, but they work very quickly and can analyze millions of sources of information in a matter of seconds and quickly categorize them.AI RECRUITMENT SOFTWARE

Using statistical data, AI systems are able to "predict" and "learn" by constructing possible decision curves and then optimizing the decisions based on multiple criteria. Therefore, it is not difficult to imagine an AI system that considers all possible demographics, work experience, and interview questions with candidates, and then “predicts” how effectively each of them will perform their job (HiredScore, Pymetrics, HireVue, IBM and other companies already are working on it).

Despite the fact that the process itself is much more complicated than it seems, the solution of this problem is an important and noble cause. Answering a question on this topic a few weeks ago, I noted that “most management decisions are made by us today exclusively on an intuitive level. If systems like this make us a little smarter, then we can significantly improve our operational efficiency.”  

Of course, there are a large number of risks and obstacles that still need to be overcome, but the potential is simply recruitment software

What kind of applications can be expected in the near future?

Let me list just a few of the areas of great potential.

In the field of recruiting,  many decisions are made intuitively. One study found that most hiring managers make judgments about a candidate within the first 60 seconds of a meeting, often based on the candidate's appearance, handshake, attire, or speech. Do we know what characteristics, experience, education and personality traits guarantee success in the performance of a particular role? No, we don't. Managers and HR professionals spend billions of dollars developing assessments, tests, simulations, and recruiting games, yet many claims that despite this, 30 to 40 percent of the time, candidates are selected incorrectly.

AI-powered algorithms can scour resumes, find suitable candidates within companies, identify high-performing employees, and even provide interview transcripts, helping us choose the talent most likely to be the most successful. One of our clients uses Pymetrics AI-powered gamification assessment to screen applicants for marketing and sales jobs. Eliminating all of the mistakes made in the interview process and reviewing the "track record" of candidates made in the current process, the success rate has increased by more than 30%. AI in recruitment has a great future.AI RECRUITMENT SOFTWARE

It should also be taken into account that despite the general concern about professional skills (software skills, sales skills, math skills, etc.), most research shows that technical skills are only a small percentage of success. Most of the recent research on high-performing recruitment processes shows that Maturity 4 companies, i.e. those with the highest financial performance due to competent recruitment, rely (40% of recruitment criteria) on emotional and psychological characteristics, such as as ambition, learning, dedication and purposefulness. Will it take AI into account? Perhaps.

(Suppliers in this market include companies such as LinkedIn, Pymetrics, Entelo, HiredScore, IBM, Textio, Talview, Unitive, PredictiveHire, and more.)

In the field of personnel development and training,  we really do not know how to "train" employees. More than $200 billion has been spent on the global learning and development industry. However, most training experts say that at least half of these funds were wasted (the solutions developed are forgotten, applied inappropriately, or simply a waste of time). However, we still do not fully understand which half. 

Do you know what you “need to learn” in order to perform better? We can only guess, what if we had algorithms that could track and learn the knowledge, behavior, and actions of our team's top performers and then simply explain what is needed to match them? Similar algorithms like Netflix are already being used in the learning platform space, making learning as rewarding and fun as watching cable TV. The market is still young, but there are plenty of opportunities. Our research shows that, on average, employees have less than 25 minutes a week to train, but if that time is put to better use, each employee's effectiveness will increase.


(Suppliers in this market include Degreed, EdCast, Filtered, Volley, Axonify, BetterUp, Clustree, Workday, and more.)

In the field of management and leadership,  we act like Zen masters. We read books, attend seminars, copy the leaders we admire, and praise the successful leaders of today. Do we own the science of management? I guess we rarely think about it. Today we focus on purpose, mission and commitment. Just a few years ago, there was "service leadership", and in my youth, "performance and financial sense" were valued. Most research shows that there are dozens of management and leadership traits that determine success, and each of us offers a unique combination.  

AI can help us identify these features. I know three vendors who have created AI-powered learning tools and systems that solicit feedback, read comments, and read the mood of employees and teams. They use this data to compare personal and team results with those of higher-performing teams, allowing managers and supervisors to understand what is needed to get the job done more productively. One of my clients said that in just three months of using this tool, the effectiveness of the company's management in terms of strengthening corporate values ​​increased by 25% thanks to only small behavioral features.

(Suppliers in this area include companies such as Reflectiv, BetterWorks, Ultimate Software, Zugata, Humanyze, ADP, Impraise, and others.)ai recruitment software

 There are also great opportunities in the areas of misconduct and compliance . The results of one study showed that employees who steal or commit crimes negatively affect everyone else (other employees begin to copy this form of behavior). AI can look at organizational network data (emails, comments) and identify stress areas, possible ethical violations, and many other forms of compliance risk, as well as highlight “red zones” for HR or compliance directors to they had the opportunity to intervene to prevent dishonest practices.

(Suppliers in this area include TrustSphere, Keencorp, Volley, Cornerstone, etc.)

In terms of employee well-being and engagement,  AI is used to identify behaviors that contribute to performance degradation. In the field of safety, AI is able to identify behavioral factors that lead to accidents. New analytics tools can identify signs of stress and misconduct and alert HR or line management.

(Suppliers in this area include Limeaid, VirginPulse, Glint, Ultimate Software, CultureAmp, TinyPulse, Peakon, and more.)

In the area of ​​employee self-service and candidate management, new smart chatbots make it easier and more efficient to interact.

(Suppliers in this area include companies such as IBM, ServiceNow, Xor, Mya, Ideal, Paradox, and others.)

This list is endless.AI RECRUITMENT SOFTWARE

Are there risks? What's going on with HR analytics?

All of these apps are new, and while attractive, they come with a lot of risks to consider. The main risk is that AI cannot work without "training data". In other words, algorithms use the experience of the past. If your current management practices are biased, discriminatory, or overly hierarchical, then you can only make matters worse. We need an "objective" AI that we can "customize" and whose algorithms we can control to make them work effectively. As with early cars that didn't always run smoothly, our first developed algorithms require "bumpers" and "sticks" so we can make them more accurate

Systems can help perpetuate bias. Let's say that your company has never hired a woman as an engineer and that you only have a few African American engineers on your staff. Clearly, an AI-powered recruiting system will decide that promoting women and black engineers to leadership positions is wholly undesirable. This type of bias should be eliminated from the algorithms, but this will take time.

There is a risk of disclosure as well as unintentional misuse of data. Take, for example, the use of analytics to determine the likelihood of a high-performing employee leaving a company. If we tell management that “this employee is most likely going to leave”, then we can form the wrong behavior - management will begin to ignore this employee or change their attitude towards him. We must learn how to apply behavioral economics correctly so we don't accidentally turn AI into a HAL computer system (2000 HAL movie). Today, AI is a “tool” for making suggestions and improvements, not a system that makes decisions on its own.

I recently spoke with an AI executive at Entelo and we discussed the need for "explaining" and "transparent" AI systems. In other words, when a system makes a decision, it must tell us why it made the decision so that we (humans) can decide whether the decision is worthwhile. He told me that for his company, this is one of the most important criteria when developing new tools, but most AI systems are still built, unfortunately, in a "black box" manner.

Consider what might happen if an autonomous system fails. We will need a lot of time to analyze the current situation, understand which visual or algorithmic systems failed, and also to study what could lead to such a situation. What if the AI ​​gives the wrong recommendation about a candidate, salary adjustment, or management intervention? Can we find out? Will we be able to detect it? Will we even notice until it's too late? There is still a lot to be done to figure out how to “teach” our AI-based control systems to work properly.

Will AI become the hallmark of HR decisions?

Currently, the hype around AI is very high. Every HR software vendor wants to make you believe that their machine learning team has a best-of-breed AI solution. Of course, opportunities in this area are of great importance, but do not be influenced.  

The success of an HR tool depends on many things: the accuracy and completeness of algorithms, the ease of use of systems, but, more importantly, the ability to provide the principles of the so-called "narrow AI" (or specialized solutions that can solve your specific problems). This can only be achieved if the supplier has a large amount of data (to train the system) and receives a large amount of feedback on the results of the system. Therefore, the main difficulty, in my opinion, lies in setting directions, developing a business strategy and establishing a trusting relationship with the client, and not just having professional engineers.  AI RECRUITMENT SOFTWARE

And don't buy a black box system unless you can test it in your company first. All decisions made at the level of management or employees of the company are often based on the principles of culture, so it will take time to use the systems in real life and customize them to our needs. For example, IBM has spent years optimizing compensation and human resources solutions for its company based on its culture and business model. Now they offer their tools to corporate clients, and each implementation opens up something new for them about algorithms, helping them to optimize them according to industry characteristics, culture or organizational needs.

Despite all the difficulties, the potential is huge

Despite all these difficulties and risks, the potential is simply incredible. Companies spend 40-60% of their revenues on payroll, and most of this huge amount is the result of managerial decisions that are made only on the basis of intuition. I am confident that as workforce AI systems evolve, become more robust, and become more problem-solving, we will see significant improvements in productivity, efficiency, and worker well-being. We just have to be patient, vigilant and ready to invest in the future.

Post a Comment