IT network administrators take note: AI is not a panacea

If there’s one constant in the tech world, it’s the constant tug-of-war between hype and reality. I’ve seen this happen when a new “transformative” technology comes on the scene. With the rise of artificial intelligence (AI), it’s back to the future as we wonder how this promising advancement will change network management.

In theory, AI should be a game-changer. Network teams will be able to identify problems in real time and anticipate potential problem areas before they become critical. The same goes for tracking traffic patterns and managing network performance. The result: better use of network resources, fewer support calls and more satisfied users.

But before they jump in, network administrators should take a closer look at what an AI transition means in practice and try to separate the hype from the reality.

Take stock of your infrastructure

With increasing complexity and proliferation of devices at a record pace, network administrators’ tasks have become much more difficult. IT budgets continue to shrink and organizations seeking to reduce network support expenditures are stretched IT departments operating at dangerously thin levels.

Here, network teams can use AI to dig out a hole.

The ability to troubleshoot and resolve issues faster translates into reduced network downtime and improved network performance, while reducing overall IT costs. It also helps to give customers a great experience with fewer support calls and fewer complaints.

Here’s a real-life example of how the industry can help support that. By building AI into network solutions, technology providers can create conditions where when a customer reports a problem, they take a snapshot of the entire network and run the data through a learning engine to find out what happened.

Using an AI/ML engine that learns from issues seen on other customer networks ensures that issues once spotted are not repeated elsewhere. This saves a huge amount of time, as problems can pop up anywhere. A failure may be related to software being loaded on an access point. Or maybe it’s in the support network. But with the help of AI, an organization can now get a detailed picture of what’s going on in a fraction of the time it used to take to solve the problem.

Unlock big data

AI will be especially useful when it comes to parsing the massive amount of client telemetry generated by a network infrastructure. In the past, the only way to extract information from all this data was with the help of (well-paid) experts who knew their way around different network technologies. However, if a company couldn’t afford the right staff, this treasure trove of valuable data was largely underused. This is especially amplified when customers deploy network solutions from different vendors, preventing a single view of the network.

Using AI tools, organizations can now solve this big data problem and get the insights they need to address questions facing IT departments, including:

• Which sites and clients are experiencing poor network experience?
• What are the root causes of poor performance?
• Which sites are running at full capacity and what network changes are needed to improve the situation?
• Can the network be continuously scanned automatically to maintain a good security status for the network?
• Do IoT devices present security vulnerabilities?
• Do network services work well during peak times in my network?

Don’t get caught up in the hype

There is no doubt that AI is becoming increasingly relevant for network management. And as processing power increases, the technology will continue to get better. But be smart about how you use it. Don’t ignore the fact that AI is not something to be applied blindly.

Some mundane and manual tasks are still better automated. For example, you don’t need AI to issue network patches. That’s why I think not everything can or should be handed over to AI, which can get expensive when you implement these kinds of solutions.

Focus on your use case. What business problem are you trying to solve? This may seem rudimentary, but all too often this basic question is ignored.

Second, does it fit your economy? Every business must stick to a budget. Make sure any AI implementation doesn’t break the bank.

Third, test it out to make sure the network you’ve deployed really delivers the results you want. Does it help you solve the business problem? How does it do that? And does it work reliably?

Choose pragmatically, not based on hypes

There are a number of tools out there. Some of them are AI powered, and some are not. Don’t get caught up in the hype. Instead, make sure you pick the ones that solve your problem. Otherwise, your expenses will only increase exponentially.

Above all, keep in mind that this AI transformation will not happen overnight. Throughout my career, I’ve seen this happen every few years as markets strike the right balance between enthusiasm and excessive confidence in new technology. This is all exciting, but plan your trip step by step.

As AI begins to gain confidence and more automation takes place in the network, you can expand your capabilities accordingly. This is a journey that will take time, and your patience will pay off. So, take it step by step.

Rad Sethuraman is VP of Product Management at Cambium Networks.