No One Wants to Talk to You: Cherwell Exec on How AI Is Changing Service Management

Written by Kris Blackmon
November 10, 2017

Steve Rodda of Cherwell Software explains why AI and automation aren't going to put service techs out of a job. Instead, they'll increase IT's value proposition and make partners even more indispensable than they are today.

Gartner predicts that 85 percent of all customer interactions will be managed without a human in just four years. When it comes to the day-to-day operations of an IT helpdesk, that’s going to mean a lot of changes in how level 1 service techs spend the majority of their time.

Colorado-based Cherwell Software is hoping to be at the vanguard of using AI to automate the mundane tasks that take up so much of IT technicians’ time. The company recently landed a $50 million round of investment from KKR, and it has big plans to use those funds to modernize the helpdesk. Steve Rodda, Chief Product Officer at Cherwell, says the type of automated solutions his team is developing are going to be key to equipping IT departments and channel partners to assume broader, more business-focused roles.

Unlike sectors such as manufacturing where AI and automation may very well mean the loss of jobs, Rodda says that these technologies will increase the value proposition IT workers provide an organization and fundamentally change the way IT service is delivered.

But we’re still in the nascent stage of these technologies, and there are a lot of business best practice kinks that need to be worked out before mainstream adoption is a reasonable possibility. Rodda compares it to the self-service portal craze that consumed IT a decade ago. In theory, he says, the portals were there to do the exact same thing Cherwell is hoping to accomplish with AI: eliminate the mundane tasks of level 1 IT service techs to free up their time for more business critical functions.

“But we really didn’t get there because it’s hard to automate across other systems and access and security and those kinds of things, but AI will help bridge those gaps. Not only will it be able to look up the knowledge base article that is there, but it will be able to look at what the last 10 technicians tried, and then work through some of those mundane troubleshooting tasks.”

If AI can’t solve the problem, Rodda says it will still end up in a tech’s hands, so the need for IT support isn’t going to go away. There’s just going to be some precursor work that will be done ahead of time by an AI system that will result in higher and faster close rates, but not the elimination of overall jobs.

The kicker is that it can’t just be AI if the business world truly wants to leverage cognitive solutions to improve business processes. Rodda says that AI hasn’t broken through today because it takes of inordinate amount of training to actually get the systems up to speed. Some of Cherwell’s vendors estimate that for some tasks, it can take AI systems up to 80 days before they actually can learn all the different and various ways to solve a task before and actually automate it.

“That barrier of entry or time to value for AI is not there yet, which prevents a lot of adoptions. You need the combination of a good NLP and AI system, then you need a system in place where the AI can actually learn from the normal, everyday tasks the technicians have in place.”

At Cherwell, one of the ways the company is attempting to work around this challenge is by allowing the technician to run commands through their normal interface and having the AI system simply listen and learn from the system of record. Training then becomes not an independent process to make the system smarter, but rather a cognitive computing process that learns from technicians’ normal interactions.

The reluctance to integrate AI into existing support systems Cherwell sees from traditional VARs and MSPs isn’t surprising to Rodda.

“This value proposition as it’s presented to MSPs and channel partners is that it’ll essentially put them out of a job. But reality of the situation is that these things are light years away from being able to replace an MSP or VAR in that type of sense. What it will do, probably in our lifetimes, is make the tools they use better today. In the MSP world, you already have automated tasks that can fire off an event that happens and it gets taken care of if whether it’s in your RMM solution or whatever the case is. In the ITSM world, we have automated tasks driven from the DevOps culture that automate these basic tasks.”

In the short term, Rodda says what partners need to be worried about is finding the tools that aren’t trying to sell them on the hype of AI, but will simply put these technologies in place in the background to make processes more efficient. For example, an MSP that’s already switched from break-fix model to a managed services model will be better enabled to get involved in solving business problems and achieving business solutions.

This is the “heart and soul” of why Rodda doesn’t think these technologies will actually put MSPs or IT departments out of business. “There’s no other place that has visibility across the entire organization except IT—and maybe the CEO, but he doesn’t count,” he laughs. “And no one else is responsible for looking after the entire business question.”