Artificial intelligence can provide timely and efficient IT support, freeing up tech pros’ workloads.
In order to plug AI into your existing workflow you must first understand and organize master data sets, says Schneider Electric Chief Digital Officer Herve Coureil.
As a system administrator working in technology for more than 25 years, I’ve provided some measure of IT support. While it’s rewarding to help people, sometimes the monotony in doing so can lead to feeling unfulfilled, especially when there are better and more promising tasks afoot—if you can just find time to work on them.
It’s equally challenging for users and administrators to handle both support ticket platforms and the communication gaps and lags, which can hinder or impede resolutions, plunging both sides into bureaucratic red tape that lowers satisfaction all around as unresolved issues pile up.
SEE: Managing AI and ML in the enterprise 2019: Tech leaders expect more difficulty than previous IT projects (TechRepublic Premium)
This is where artificial intelligence (AI) can play a role. AI can help to free up technology pros for more meaningful endeavors in a more streamlined, goal-driven fashion. To learn more, I spoke with Bhavin Shah, CEO and co-founder of MoveWorks, an AI platform focused on providing support solutions.
Scott Matteson: What’s holding back IT support today?
Bhavin Shah: Today, the average IT support ticket takes three calendar days to resolve. This process is painfully slow largely because it’s managed by a long chain of people. Also, employees are very good at describing their problem — whether it’s resetting a password, unlocking an account, getting a license for an application, or getting an answer to a simple question — but they don’t always know what system to access to resolve the issue.
We have IT help desk people that handle this type of diagnosis. But IT teams still spend far too much time working on highly repetitive tasks when they’d rather focus on more strategic parts of the business.
Scott Matteson: Why has IT support made such little progress in the last two decades?
Bhavin Shah: On the backend, IT support has made some progress. It’s at least looking better than it did 10 years ago. But not nearly enough has changed. The reason is that most IT ticketing and portal systems are unable to make the direct connection between what the employee needs and the mechanism to trigger the resolution. Current solutions focus on providing workflows to route and manage ticket queues, leaving the actual work of interpretation and understanding to IT service desk agents. It’s also because the systems deployed today don’t actually do the work to resolve the issue.
AI as a solution
Scott Matteson: How can AI ameliorate the bottleneck in IT ticket resolution?
Bhavin Shah: AI has the ability to accurately understand how employees describe their IT issues. It can quickly figure out a task, and execute it instantly. It can learn and improve over time from hundreds of millions of data points and become more accurate — a task that would require a lot of time from an IT team.
It can also save IT departments money. An average employee submits one IT support tickets per month and, on average, each IT ticket costs about $25 to resolve using traditional methods — e.g., service desk agents, workflow tools, service centers, etc. Our average customer has more 10,000 employees. That’s a lot of money for any size company.
AI enables IT teams to resolve these issues autonomously, and at a fraction of the cost, with much less hands-on human intervention. And that reduction in overhead starts from day one.
Scott Matteson: What are some common examples of problems which have AI solutions?
Bhavin Shah: Our machine learning (ML) models have analyzed tens of millions of IT tickets. The most common issues we resolve include requests for access to software, account unlocks and password resets, creating or adding users to email distribution lists, providing steps for software troubleshooting, and completing hardware requests (laptops, peripherals, etc.).
We currently resolve 20-35% of issues for current customers—a year ago this was about 10-15%, which shows our continuous learning system has become smarter and more efficient over time.