STIR/SHAKEN gives carriers ways to protect consumers and businesses from robocalls. However, the technology has some loopholes that make it less than ideal. Until regulators and industry leaders can close those loopholes, carriers will continue to rely on analytics engines (AEs) to help them identify scam calls before they reach consumers.

Not sure what a carrier analytics engine is? The following explains how some of the most effective AEs work.

Carrier analytics engines use several strategies to identify and combat robocalls. Some of the most useful approaches rely on machine learning, behavioral analytics, and audio fingerprinting.

Each of these options offers a unique approach to spotting and fighting scams. Some can even work together to pinpoint suspicious activity.

Machine learning makes it possible for algorithms to review massive amounts of information to identify suspicious patterns. Machine learning analytics engines look at metadata such as:

The industry has more than enough metadata to accurately identify robocall patterns. Perhaps even more importantly, machine learning algorithms evolve over time. Scammers and spammers will inevitably look for new ways to avoid detection. Machine learning AEs pay attention to emerging trends to identify those new strategies as quickly as possible.

Analytics engines often look for behaviors associated with spam and scam calls. For example, a behavioral analytics engine might look for:

These and other types of information are saved in carrier and third-party databases. Analytics engines can then review the stored data and apply labels to activities that appear similar to those used by scammers.

Audio fingerprinting is a relatively new and interesting approach to identifying scam-likely calls. Audio fingerprinting technology involves detecting speech, examining the speech content, adding the information to a database, and using algorithms to determine whether the speech conforms to patterns associated with spam and scam calls.

Audio fingerprinting algorithms make decisions based on user feedback and machine learning. This combination of technologies allows companies like Robokiller to compare call content and find call scripts.

Obviously, many legitimate companies use call scripts to help agents connect with customers, leads, and prospects. However, robocallers often use prerecorded scripts that sound exactly the same each time. Even when a scammer uses hundreds of phone numbers to evade detection, audio fingerprinting quickly recognizes the script and labels calls to protect consumers.

Similarly, audio fingerprinting helps prevent number and brand spoofing by focusing on call content instead of just phone numbers.

Major carriers in the U.S. usually partner with third-party companies that develop carrier analytics engines. For example:

Since carriers use different AEs, labels can differ from company to company. The varying algorithms and labels also make it important for you to use call monitoring that works across all major carriers.

Carrier analytics engines are some of the most effective tools in the fight against robocalls. Algorithms don’t work perfectly, though. There’s always a chance that they will apply negative call labels to numbers used by legitimate businesses. That means you should take a proactive approach to monitoring your number reputations and remediating flagged numbers.

Protect your numbers from inaccurate flags by:

BatchDialer has a suite of products developed to keep your phone numbers’ reputations pristine. Phone number monitoring shows you when your numbers get flagged, so you can address the situation immediately. Actual device testing uses real devices to show you what caller ID screens show when you call them. You can even get help with number redress remediation to get your numbers back in rotation as soon as possible.

Sign up for a five-day free trial to learn more about how BatchDialer benefits call centers and other organizations that make a lot of outbound calls.