Integrating AI & ML Into Business Analytics Applications

NETSCOUT “NO LATENCY” podcast discusses two of today’s hottest tech topics.

Integrating AI and ML into business-analytics applications

NETSCOUT Senior Director of Engineering Greg Mayo recently joined AVP of Global Services Operations Steve Sviontek on our “NO LATENCY” podcast to discuss two of the hottest topics in the modern world: artificial intelligence (AI) and machine learning (ML). Mayo joined NETSCOUT in 1998 and has worked in several areas of the business over the past 25 years. Beginning with a 10-year stint in engineering, he then transitioned to the CTO’s office in 2008 to focus mainly on telecommunications; since 2020 he has been a leading member of the nGenius Business Analytics (nBA) platform team, where he works to provide automation capabilities for customers.

Beyond the Buzzwords

Since the launch of ChatGPT in November 2022, AI and ML have been among the most common buzzwords in the world—perhaps outside of Taylor Swift and Travis Kelce. In reality, AI/ML has been around for years, even decades. They are currently seeing a resurgence in popularity with large language models (LLMs) such as ChatGPT, Google Bard, and others. In fact, NETSCOUT has been using AI/ML techniques in our products for many years.

According to Mayo, AI and ML are, simply put, collections of algorithms that are skilled in specific areas, working together to form the complex AI models we know today. Some of these algorithms are designed to pick out specific elements of an image, understand language, recognize patterns, provide automation capabilities, and perform other useful tasks. To some degree, all NETSCOUT products have an element of AI/ML integrated into them.


With the problems NETSCOUT solves, it is challenging to fully automate issue resolution with AI. Although current LLMs, such as ChatGPT, can pass the LSAT, MCAT, and other exams, they struggle to solve the complex, nuanced problems NETSCOUT customers face. That is because you can find all the answers to the exams listed in the set of resources used to train the language model—it’s black and white. Performance problems are gray: They don’t have a definitive answer, because the issue in question could be caused by a multitude of problems. There is no “book” to point directly at an answer. That said, NETSCOUT products use a series of complex algorithms to help identify the cause of the issue based on patterns and consistencies in what the products have already learned and apply the correct fix to the problem. This expedites mean time to resolution (MTTR), maintains user experience, identifies security threats, and mitigates availability issues such as distributed denial-of-service (DDoS) attacks.

Mayo explains that NETSCOUT feeds its AI algorithms with proprietary Adaptive Service Intelligence (ASI), which turns wire data into what we call Smart Data via deep packet inspection (DPI), as a data source. The detailed dataset allows the AI to do its job more efficiently and effectively, aiding in the ability to identify and troubleshoot performance, security, and availability issues based on past trends. This allows the AI algorithms to continuously learn and grow, getting more and more proficient at the tasks they are assigned to complete.

NO LATENCY Podcast Series: What is all the Fuss About ChatGPT?

NO LATENCY Podcast Series: What is all the Fuss About ChatGPT?


Learn more about nGenius Business Analytics.