The advancement of technology over the past decade, particularly in artificial intelligence (AI), has created incredible new opportunities for managed services providers (MSPs) that help support clients with their technology needs, as well as the potential to leverage these technologies for their use.
AI and machine learning are a broad category where computers are taught to think like humans and perform tasks on their own, or even advance further to develop their insights on situations. This is often accomplished through algorithms and training the computers on large data sets to model the human brain.
One subset of the AI and machine learning category is deep learning, a type of machine learning through which computers learn to perform tasks by looking at large data sets, such as images, text, or sound. The data sets are typically labeled to guide the computer on the correct outcomes and then through additional neural network architectures, which are complex structures that mimic the human brain by taking in many inputs and converting those into a single output.
Over time, these outputs increase as increased layers are added to the algorithm. For example, early instances in images might be able to identify the outlines of different figures, while additional layers over time with increased learning can identify letters, faces, animals, or other items. These multiple layers are where the name “deep learning” comes from. Building and training these models is often highly involved and intensive, requiring greater hardware capabilities to handle the processing. However, as these deep learning capabilities increase with additional training and time, the result is new capabilities that were not previously possible.
Further reading Key Reasons to Reconsider Your Endpoint Protection
Deep learning is currently in use in several use cases. If you use voice control for your TV or a hands-free speaker in your home, you’re likely using deep learning technology without realizing it. Additionally, it is the technology behind driverless cars, teaching them how to navigate city and rural streets and follow the rules of the road, including stopping at stop signs and recognizing pedestrians. These are just a few examples of deep learning at work within our regular lives today. Additional use cases can be found in computer vision, speech recognition, translation, biometrics, drug design, medical and other image analysis, inspections, climate change presentation and analysis, and more.
These use cases have in common the need for computers to learn by example, which is exactly what deep learning accomplishes. We have perhaps just scratched the surface regarding applications of this type of technology, but it provides incredible promise for the technology industry at large — a benefit that will trickle down to MSPs to improve their ability to support their clients on many different fronts.
Cybersecurity applications for deep learning
Cybersecurity has become one of the fastest-growing value-add categories for MSPs to offer to their clients over the past few years, thanks to a seemingly never-ending increase in cyberattacks worldwide. According to one survey, cybersecurity represents a $2 trillion opportunity for cybersecurity technology and service providers — including MSPs.
There are many cybersecurity technologies that an MSP can leverage to build out its cybersecurity practice, from patch management to two-factor authentication, to endpoint prevention and protection. Deep learning presents the opportunity to grow this set of capabilities for an MSP and increase its ability to identify potential threats, as well as mitigate them.
One way deep learning can help is to improve intrusion detection and prevention systems, which detect malicious activity on the network and alert administrators to investigate and remediate if necessary. Often these products work by having a list of patterns or signatures of known attacks, then looking for those signals throughout the organization. Deep learning could help refine this process, helping analyze traffic more accurately to limit false positives and potentially identify more sophisticated attacks. Similar capabilities can also be applied to network traffic analysis to detect potential cyberattacks, such as SQL injections and Denial of Service attacks.
Another place where deep learning can help is in the area of spam and social engineering. Through a deep learning technique called natural language processing (NLP), IT administrators could more thoroughly detect spam emails and phishing attempts that are pretending to be something they are not, in order to deliver malware to the organization.
Finally, an additional area where deep learning could be applied within cybersecurity is user behavior analytics. In this technology, IT and cyber teams look to track and analyze user behavior and activities to establish a baseline of normal for the organization, then flagging any potential anomalous behavior from that baseline as a sign of potential attack. This helps organizations pinpoint insider threats and malware that has bypassed initial cybersecurity protections.
These tools have in common that each can benefit greatly from an increased ability to learn how attackers move and act as they try to target the organization. As these tactics change over time, leveraging deep learning could allow for an increased ability to identify and adapt to these new tactics and protect and defend the organization.
Managed services providers going deep (learning)
For MSPs, these technological enhancements add significant benefits to their ability to service their customers with the latest technologies and capabilities. While it is unlikely — at least at this current stage of technology advancement — that most MSPs will be developing deep learning technology algorithms themselves, they certainly can benefit from the advancements at hand as they look to grow their business.
MSPs are already beginning to see the fruits of these technological advancements. While MSPs may not leverage driverless cars or voice recognition, the benefits to cybersecurity tools previously described are a great example of how deep learning can improve an MSP’s capabilities through the tools it may already use to better support its customers, particularly in an area such as cybersecurity that is in high demand and commands high margins for MSPs.
Deep learning can also help support and refine automation for MSPs, as many increasingly automate their environments to enhance efficiencies and respond more quickly to customer challenges. Deep learning technologies incorporated into the tools an MSP relies on to complete these daily tasks could help support these goals.
While deep learning is arguably in its infancy as a technology, it is clear that it holds incredible promise as it continues to develop and the understanding of its potential use cases is more clearly understood. MSPs should watch this space carefully to ensure that they are taking advantage of the latest cutting-edge technologies and leveraging the benefits for customers. Just as with any new technology areas that emerge, MSPs that keep themselves abreast of the latest technologies and incorporate them into their strategies as and when it makes sense to do so will be more likely to stay a step ahead of the competition and set themselves up for long-term growth.
Assess vulnerabilities and threats, network security, workspace and equipment security, documentation, and more. The pack includes:
- a ready-to-print PDF file
- an Excel file to help create a customizable assessment resource
About the author
Kurt Abrahams is the Vice President of Marketing at MSP360 with expertise in technology marketing, cybersecurity and AI based technology.
More articles by Kurt Abrahams
