95% of large companies use automated hiring systems. Andrey Krylov, founder and CEO of the Hiberty personnel hiring system has put together five common mistakes made that prevent businesses from making the most of automatic recruitment.

According to LinkedIn, over the past five years, the number of HR specialists with data analysis skills has risen by 242%.

Why do recruiters re-qualify as analysts? The trend is connected to the appearance of new digital hiring tools. Around 5-7 years ago, large companies like trading networks, banks, and mobile operators spent hundreds of hours developing forms and communication with jobseekers. Rank-and-file personnel are constantly being updated, new people needed to be hired every day, but the vast majority of these tasks had to be done by hand.

A colossal amount of time was spent on digital bureaucracy.

Recruiters at companies like Pyaterochka, Samolet, and Perekrestok spend 40-50% less time on processing information, and rank-and-file positions are closed within just 3-4 days.

The process of automating hiring is done all over the world. For example, Amazon has practically removed recruiters altogether from the process of hiring warehouse employees. To start work, all an employee needs to do is fill out an application, watch a training video, and take a few tests.

But how effective are these services? Of course, they save time and money for HR departments, as well as reduce downtime and make it possible close urgent vacancies faster. However, it’s important to take into account the potential complications when selecting automated systems.

For example, research conducted by Harvard Business School has shown that AI based systems in the United States screen out up to 27 million job hunters without giving them the chance to sit for any interviews. Problems usually appear due to poor quality software, or the company’s approach to auto recruiting. Let’s look at what mistakes employers make and how to avoid them.

1. Ignoring customization

Automated hiring systems are by themselves quite neutral. The technology allows you to process big data in a short period of time, which definitely makes it possible to save time. But a lot depends on the parameters and settings set by the developer or the employer themselves.

Smart systems screen out millions of candidates due to excessively formal parameters. For example, if a person hasn’t worked for 6 months because they had to take an extended break for health reasons, but the AI system doesn’t recognize the reason, then it might consider it simply a long-term employment gap. The job hunter will immediately be marked as unreliable and will be screened out from the selection.

Another example is excessively strict age boundaries. A candidate who is a couple of years older than the age bracket may have an ideal set of skills but the system simply won’t recognize it. That’s why it’s important when selecting an auto-recruiting system to specify what filters the company offers, how often they’re updated, and whether they can be customized.

For example, we initially used a simple resume keyword search, but we soon developed a system that would search through individual lines. This helps avoid confusion, and potentially highlight a person who knows English well, rather than someone who worked at reception at an English school.

These details can significantly affect output. And when the company is closing rank-and-file vacancies, then even saving just a few seconds without losing efficiency can be of use.

What to do?

Give preference to vendors who use up-to-date text recognition systems and regularly update screening technologies and models. One 2020s trend is to create a digital profile of an ideal candidate based on the best workers.

The system takes this benchmark into account and searches for someone who is a good match based on skills, experience, education, and other parameters. This technology is often used in banking.

2. Not trying to build an ecosystem

The candidate selection algorithm needs to be built into the existing infrastructure, otherwise it won’t be of sufficient use. Using only auto-recruiting in isolation from the rest of the modules, means you will gain an extensive database of formally suitable candidates, but in the end, you will still get stuck in the digital bureaucracy.

Before introducing AI into recruiting, try and determine which components to use to unify the system. This could be a staff administration system, a candidate checking system, a call center for processing applications, or a system for remote testing and training for candidates.

What to do?

When talking to developers, describe the hiring pipeline that the candidate follows and make sure the company knows how to set up integration.

X5 Retail Group is an example of a company that uses an ecosystem-based approach. Previously, a recruiter called the candidate, and then a long approvals process began, sometimes the potential employee didn’t wait for a final decision and found a job somewhere else.

Hiberty automated part of the processes and reduced waiting time, which meant that 34% of people were able to be hired on the day they applied. But this required vacancies to be ranked. Generally, salespeople can be hired more quickly, but more preparation is needed for supervisors and directors. This was previously all taken into account by the recruiter, but now the platform deals with it.

Try and choose omni-channel services, that don’t just give you a database of suitable candidates, but also help with registering and onboarding for new employees. Remember that your goal isn’t just to find a person for a position as quickly as possible, you also need to find a relevant candidate and at the same time keep recruiters out of routine tasks that an algorithm can deal with.

So try and think through how best to optimize the whole recruiting process at all stages using all available tools.

3. Forgetting about people who can “break into” the system

Reddit and other forums have plenty of threads where users discuss “hacking” smart recruiting systems. And there was recently a story in the media about how one job hunter intentionally embellished her resume with fictional skills and ridiculous jokes, but also mentioned that she worked at LinkedIn and Microsoft.

Despite all the absurdity, recruiters called her in for interviews, and at top companies at that. It’s clear that at the first stage of recruitment, candidates can adapt their resumes to fit algorithms and get invitations from companies. They likely won’t be hired, but it will waste the company’s time.

What to do?

Full automation of recruiting still represents a large risk. We recommend that companies involve an HR manager at some stage.

For example, if the system finds some kind of deviation, then the resume can go for manual validation with a recruiter. A quick suitability check can be done by a bot.

In any case, there will always be attempts to deceive the system, but they are a small percentage; in most cases it’ll be people who are genuinely looking for work rather than trying to deceive the automated selection system.

4. Relying on mass hiring without a properly developed HR brand

Mass hiring practices may scare off jobseekers who evaluate companies not just by pay and benefits, but also by culture and values.

Auto recruiting often causes negative feedback from IT specialists who get automatic e-mails from companies every day. This often causes conflict between the two camps. However, recruiters themselves recognize that they don’t fully understand IT and send out mass offer e-mails to at the very least close some vacancies in the overheated labor market.

What to do?

First, you need to decide which selection mechanism you want to use for different candidates. For example, it’s better to search for data science experts with a rare set of talents manually, whereas junior testers can be selected automatically.

If you have typical questions and tests during interviews, and skills can be standardized easily, then don’t be afraid to use AI. If the vacancy involves in-depth screening for candidates with management skills, analytical thinking and working in a team, then in-person interviews and traditional questionnaires are the way to go.

Overall, it’s more appropriate to use automated tools if you’re looking for more rank-and-file employees in segments such as retail, logistics, and transport.

Nevertheless, companies ought to think about developing their HR brand, it’s important to demonstrate that you’re not just filling your team, but also care that you care about your employees’ development and provide opportunities for career growth in any position.

A solution can involve additional hiring practices, such as events, bonus programs, and careers fairs, as well as corporate training.

Details that seem small at first glance can strengthen your brand considerably. For example, creating a branded account in WhatsApp to communicate with candidates. You can use it to send automated notifications and launch chat bots.

5. Not having post analytics

If a company can patch holes with automated systems but doesn’t solve other recruiting and personnel development problems, then the technology won’t be of use in the long run.

Without dialog and feedback, it’s impossible to develop convenient processes and technologies for users, and in our case, candidates and employees, so analyzing this information is part of a program of constant improvements.

What to do?

Analytics can be used both for optimizing the hiring process itself, such as determining inefficient areas that use up resources. Or for evaluating employees after they start.

Developing automated solutions

Smart recruiting systems can solve issues with hiring in the long term, such as by creating a personnel reserve. This is the approach we used when helping SberStream hire developers.

First they gave quizzes, and then conducted interviews with expert recruiters. Then each candidate was given a combined profile with all their results. Those who were successful went into the personnel pool, if someone from the Sberbank team had a suitable vacancy, then a candidate could be brought in from the reserve.

The best HR practices don’t include full recruiting automation, technological solutions help with selecting candidates, but the final decision should stay with personnel department specialists.

HR technologies can save recruiters from having to handle routine tasks and free up time for making strategic decisions. Companies should pay attention to custom services so they can avoid screening out candidates because something is set up incorrectly, or the software itself is of poor quality.