AI Support Bots Still Need That Human Touch

By: Larry Lunetta, Vice President of Portfolio Solutions Marketing at Aruba, a Hewlett Packard Enterprise company

Are bots all you need? The world is in a lather about how AI will replace all things human, including network support. While AI can be very helpful in eliminating mundane, tedious tasks of data gathering and triage and even some simple network troubleshooting, it is not a substitute for personal and bespoke support services. Even with state-of-the-art data science, solely relying on AI bots for Tier 1 and Tier 2 support will often lead to endless phone trees, long lead times, wasted time and effort, unsolved issues, and worst of all–disgruntled, unsatisfied customers, employees, and partners.

Simply put, organizations should not have to deal exclusively with bots for support as a frontline strategy.

Mission Critical Networks Need Mission Critical Support

Before organizations even get to actual network operations, there is planning, design, and deployment. Here, documentation is the foundation of great support. Verified reference designs and comprehensive user guides go a long way towards smooth network operation; they are by-products of a mature, proven network solution. Often, startups don’t have them. It’s a matter of cost and blind allegiance to the marketing claims for things like AI. And, when the network has an issue, it’s likely there is just an AI bot available to help, whereas a mission-critical networking vendor will have already anticipated and solved the kind of front-end problems that only experience with tens of thousands of customers will provide. Without this foundation, even the best AI systems will fail.

At the core of providing effective support for critical issues is personalized, expedited service. In contrast to the wholesale outsourcing of frontline support to bots with little documentation, by combining best practices, best-of-breed technology and a trained staff of experts, this hybrid approach offers the best option for delivering and maintaining mission-critical networks. When a network administrator has an issue that is beyond the automated self-healing functions, the first call should be readily available and start with a dedicated support expert, who will know exactly how the network is configured, its history, and they should have all of the pertinent incident data at the proverbial fingertips. Issues can then be quickly resolved and in the event that something unexpected pops up, it can be handled without having to start all over again.

Dealing with Persistent Network Issues

Today, networks are being stressed like never before with remote work, IoT, cloud migration, and so forth, spawning novel, unforeseen issues that cannot be handled by limited AI-based tools. During these periods, accessing an engineer with intimate knowledge of the system and configuration on-site will be the lifeline network teams need to help diagnose and resolve these types of challenges.

Is this simply a luddite view of Artificial Intelligence? Why can’t network administrators simply outsource support to an AI system? There are two parts to the answer. First, while AI is a terrific aid to troubleshooting, it will not catch every issue—amazingly, even the most aggressive AI bot proponents claim to handle only two out of three issues. Combine this with the maelstrom of network change in mission-critical circumstances and you have what network engineers call the “2am problem,” when the network is failing a critical business function and the built-in support is not enough. What do they do? Does the administrator tell the CEO that some eponymous AI system was supposed to handle it but can’t? That the roulette wheel landed on one of those issues that the AI can’t fathom?

This is why a combination of integrated AI with readily accessible support engineers is the right choice. AI will get better and that will free both the network admin and support services to work on extending and optimizing network services.  But to avoid a HAL-like experience when it matters most, AI and human support must work together whenever they are needed.

An Integrated Approach to Support Services

Concerned about putting the network in the hands of unproven AI bots? Learn more about Aruba’s hybrid approach, including how it combines the best technology with personal service and engineering expertise to provide effective, around-the-clock service.

Visit Aruba Operational Support for more information.