Pinpoint Departing Users with “Predictive Flight Risk”
At Black Hat USA this year, Gurucul shared details of our most popular Machine Learning Models. Read on to learn about the second model we presented at the conference.
Gurucul Machine Learning Model: Predictive Flight Risk
How does the Predictive Flight Risk machine learning model work, what does it do? This model pulls in HR attributes such as performance review scores and the time it takes an employee to travel to work, amongst other available employee attributes. As well, this machine learning model examines relevant user behavior such as websites visited, subject lines used when writing emails, what documents a user is editing, files being downloaded and more.
For example, the Predictive Flight Risk machine learning model will detect when employees are:
- Visiting job sites like Indeed and Dice more frequently than what would likely be seen as “normal”
- Editing any documents named “resume”
- Suddenly downloading corporate secret information, such as product designs, or the like
- Sending emails with subject lines containing phrases similar to, “I’m so frustrated” or “I hate my job”, etc.
First of all, a word to the wise, if you’re working on your resume at work, STOP RIGHT NOW. It’s not a smart move and it demonstrates incredibly poor judgement. Second, even if you do hate your job, don’t send emails crying foul or complaining about your work situation. You should never send an email you wouldn’t want your boss to read. Finally, and most importantly, don’t steal information from your employer. It’s illegal, it’s bad behavior and it will definitely come back to haunt you. What goes around comes around. And, it will not end nicely for you.
Data detected by the Predictive Flight Risk machine learning model can raise a user’s overall risk score, representing the likelihood that the user is tending toward resigning from the organization. Users who are planning to resign often exhibit tell-tale predictable behaviors prior to actually submitting their resignation. It’s highly beneficial to your organization to have an early heads-up for users planning to leave so precious corporate data assets can be protected and preserved.
Use Case: Departing Users
So, someone in the organization has submitted their resignation – hurry! Let’s watch what he does! Or, consider a different path. Read on.
Instead of implementing procedures or actions within your organization to keep an eye on the activities a user who has already resigned until their termination date, you can do something vastly more effective. You can keep tabs on the online behavior of employees on a continuous basis. With this method, you’ll have the ability to predict whether they are considering resigning their position based on deviations from their regular online behavior. Behavior doesn’t lie.
Gurucul Risk AnalyticsTM employs a highly developed natural language keyword library which is applied via machine learning models to predict if a user is exhibiting behavior that would lead you to believe he or she is planning to depart. Gurucul Risk Analytics can continuously track all user’s access and activity closely to see if anyone is performing anomalous activities which may be predictive of someone dissatisfied with their job who may be preparing to resign.
Machine learning provides valuable insights into factors that can predict and reveal users who are planning to leave an organization. The prediction functionality within Gurucul Risk Analytics can enable a company to plan ahead before they face challenging skill gaps or unfortunate critical data theft.
More importantly, companies will be in a better position to take corrective action and make necessary changes to minimize issues which may be contributing to employee dissatisfaction. By bringing forth additional understanding and clarity to the data that is already available within your computing environment related to staff turnover, and by increasing the understanding of what’s going on, you can minimize employee turnover and ultimately derail resentful behavior and resignations.
What are the Benefits of Predictive Flight Risk?
This machine learning model enables companies to act before they lose critical data. We all want to stop employees from taking confidential information – especially if they are senior staff with access to sensitive Intellectual Property such as product designs, product roadmaps, merger or acquisition plans, executive salaries, and the like.
Armed with predictive intelligence pointing to potentially departing users, you can place “flight risk” users on security, network and/or application access watch lists. These lists can be tuned to place a higher level of automated access controls on employee’s who demonstrate the potential for malicious activity, especially if they are also showing signs they are planning to leave the organization.
This machine learning model can also be tuned to prevent troubled or departing users from causing damage to sensitive systems or deleting critical company data especially if their account has been granted elevated access privileges. Nothing is worse than having an irate or troubled systems administrator wipe out your entire datacenter over a difficult interaction with their boss or a bad HR review.
Gurucul Risk Analytics can assist in the detection and prevention of potential mayhem with behavior based security analytics. Contact us today for a demo or to speak with a representative.
The post Pinpoint Departing Users with “Predictive Flight Risk” appeared first on Gurucul.
*** This is a Security Bloggers Network syndicated blog from Blog – Gurucul authored by Jane Grafton. Read the original post at: https://gurucul.com/blog/predictive-flight-risk
