It’s the beginning of January, and your boss has just asked you to estimate your hiring needs for the year ahead. You’re not worried: What was once a bit of a chore is now quite easy. You just log in to your ATS and press the “Predict Hiring Needs” button. Thanks to a clever algorithm you helped program, your ATS returns a report that details how many people are likely to leave this year and which departments/domains will need extra staff. This allows you to start planning far in advance, instead of waiting for someone to jump ship.

For those of us who may be unfamiliar with the concept, an algorithm is essentially a formula that produces results based on relevant metrics. Algorithms practically run the world today. For example, Google’s success all comes down to one thing: a very, very good search algorithm.

Predicting the weather? That’s an algorithm. Modelling traffic flows at certain times of day? Also an algorithm. Algorithms are everywhere, and they are coming to an ATS near you — at least, they should be. Imagine if you could accurately predict not just the number of people who would leave your company, but which specific employees in which particular departments were likely to leave.

For example, let’s return to the hypothetical “Predict Hiring Needs” report. If you were to hit the button, perhaps your ATS would return something like this:

Churn rate this year: 14.3 percent

By department:

Sales: 18 percent

Manufacturing: 24 percent

Marketing: 8 percent

Finance: 13 percent

Perhaps you could then click on each department to view the at-risk team members — i.e., those people whom the algorithm has identified as most likely to leave based on historical data and behavioral patterns.

Think such an algorithm is just a far-off fantasy? You might be surprised at how easy building one would be. Using historical data — e.g., how many people left the company each previous year, broken down by department and job title — can already give you a reasonably accurate figure. Just add a little analysis of why people tend to leave and their average tenures at previous jobs, and presto — you have a basic algorithm that could theoretically identify at-risk employees.

If you could start each year with approximate knowledge of how many people in which roles are going to leave each department, you could plan your yearly hiring strategy in advance. While you can’t exactly interview people until a role opens up, you can start building talent pools full of people who have the skills you’re likely to need this year. As in any area of life, the more prepared you are for a worst-case scenario, the less damage that scenario will do when it happens.

Algorithms aren’t just for predicting who will leave; they can also be used to predict who will thrive at your company.

In a way, you already use an algorithm to assess a candidate’s potential; you just call that algorithm your “hiring team.” The team evaluates data on each candidate — resumes, interviews, references, test results — and determines whether or not the candidate is likely to be a good fit for the role.

Of course, the hiring team can fall short. People have unconscious biases that can sneak their way into the hiring process. But imagine if the evaluation of a candidate’s fit were largely left to an algorithm, which doesn’t suffer from the same biases: The algorithm would be more likely to rate candidates objectively, leading to more informed hiring decisions.

In the not-so-far future, interviews may disappear altogether. Artificial intelligence is getting better every day. Robots can already replace us on the production line; why not on the hiring team, too?

Why don’t more companies try to build predictive algorithms? Many simply don’t know where to start.

If you’re trying to win the war for talent — and who isn’t? — it’s time to start leveraging algorithms to help predict candidate success and employee turnover. As with the hypothetical hiring report, a hypothetical candidate success algorithm would require little more than historical data. If you look at your best employees over time, you’re sure to see some patterns emerge — e.g., top candidates scored in X percentile on this test, exhibited these personality traits, worked for a direct competitor, etc. Bad hires, too, tend to share traits — e.g., short job tenure, low salary for their level, etc.

So a predictive algorithm would be a simple game of pattern matching. Not so imposing after all, is it?

Algorithms are coming to HR. This is not a question of “if” but of “when.” Start building yours now to get ahead of the curve and see the difference it can make.

Does this mean our hard-working HR and recruiting pros will no longer be necessary once the algorithms arrive? No, but it does mean these hard-working pros will likely see their roles change to incorporate a more tech-centric focus than before.

So the next time you’re making a hire for your HR or recruiting team, you may want to plug the following variable into your algorithm: “Applicant has previous experience with algorithms = immediate hire.”

Nick Leigh-Morgan is the managing director and founder of iKrut.

in ATS]