Machine learning (ML) and the application of artificial intelligence (AI) is revolutionizing many industries. Service Management is a prime target for these new technologies and, make no mistake, they are changing the way we work in IT, and will continue to do so at an accelerating rate in the future.
ServiceNow recently acquired DX Continuum and have been working rapidly to utilize the new technology that this offered the ServiceNow platform.
I spoke with ServiceNow’s GM and VP, ITSM and Product Operations, Farrell Hough, and talked through just what the new Agent Intelligence capability brings to their customers. Agent Intelligence uses a specific algorithm that automates, with supervised machine learning, categorization, prioritization and assignment of tasks. Out of the box, Agent Intelligence applies machine learning to incidents but this can be extended to other types of tasks with little additional work.
“This is our very first step to making machine learning something that you can apply to everyday work,” said Farrell.
Agent Intelligence brings four proven areas of value to ServiceNow customers
- It is built into the platform – you can leverage the A I intelligence algorithms anywhere – This reduces the costs of applying artificial intelligence – there is no additional cost for hardware, storage or infrastructure. “AI can be very storage intensive,” explains Farrell – “but this is not and we host it in our own cloud, meaning there is no additional cost to the customer
- Agent Intelligence is easy to set up. You do not require additional purchased services to set it up or maintain. “Anyone can learn to do this, you do not need niche skill or additional training, templates can be applied against your data set quickly and easily,” says Farrell, “This is not weeks and weeks of professional services, this is just a matter of a few hours of data crunching. I know this because we did this internally and the set-up of this was confirmed”
- Time Savings – In the pilot, Agent Intelligence demonstrated an 8 percent time saving over all incidents. While the agent cannot be applied to all incidents, the application does have proven capabilities to increase efficiency and accuracy. “Some incidents are not the type of thing that you would want to predict, or they are not occurring often enough” explained Farrell. During this pilot Agent Intelligence demonstrated a 71 percent accuracy in predicting categorization and priority and it was able to be applied to 67 percent of incidents.
- Increased Customer Satisfaction – These improvements in accuracy and handling times deliver a measurable increase in customer satisfaction, due to the reduced time to get incidents in front of the right technician. Customers are also spared the frustration of having to select appropriate categories when logging their calls
“We are creating a way to be proficient that is scalable, the model can handle 10 tickets as well as it can handle 10,000. We are reducing cost, but we are also increasing customer satisfaction by making things easier and increasing the trust levels.”
Agent Intelligence does this by eliminating the triage side of incident management, something that is not visible to the customer and therefore has no impact on satisfaction” explained Farrell
I asked Farrell if she saw this new capability reducing headcount on the service desk.. “At this stage, we are not necessarily talking about reducing headcount; over a period of time that may happen, but this particular capability is about increasing capacity”.
The ServiceNow customers who were involved in the survey did not foresee a decrease in headcount in the short term. They are looking for higher quality interactions with their customers and increases in efficiency, rather than saving money on staffing costs. The expectation from these customers was that the time that could be saved by improving efficiency could be used to redeploy staff on to more fulfilling tasks.
The new skillsets that will be needed to set up and maintain this automation will create new roles to be filled, and Agent Intelligence has been designed with the expectation that it will be set up and managed in-house and will not require purchased services to maintain. But retraining will be needed to make this happen. “Sure, Agent Intelligence is going to perform a role that has been carried out by a help desk team member previously, but it requires work on the back-end to make it happen. There are skills that are needed and we will be able to tap into existing skills within organizations,” Farrell explained.
These new technologies offer huge promise to the service management industry and the people who work in it, but like every new technology, it comes with risks. I will watch this initial offering, and other machine learning initiatives, with great interest.