ServiceNow’s approach to AIOps helps IT fix system issues faster
than ever before
Whether it’s cyberattacks, software failures, or natural
disasters triggering IT issues, the end result is always the same—a
huge expense. The average cost of IT downtime runs as high as $5,600
per minute, according to Gartner. The hit to a business’s
reputation can be just as damaging.
To minimize these blows, many companies are adopting the use of AI
tools to automate IT systems monitoring, manage service health and
slash issue recovery times. We sat down with Chanda Dani, Senior
Director of product marketing, to discuss how ServiceNow empowers its
customers with AIOps.
Why are ServiceNow clients increasingly adopting AIOps?
Enterprises across the board are investing to provide compelling
customer experiences through digital services. To be competitive, they
are also constantly updating and refreshing these services. This has
led to the adoption of new architectures and an increase in the
complexity of the underlying technology behind the services. It has
also sparked an explosion in the amount of data that is being generated.
Operations teams are under pressure to meet demanding SLAs and keep
digital services running. The methods currently used to identify
potential problems take too long. In addition, whether it is finding
the root cause of a problem or fixing the problem, success depends on
the skill and experience of the person doing the job.
For this reason, the triage and remediation processes are not
scalable and are not always repeatable. In an era of digital
transformation where everybody is trying to move extremely fast, this
becomes a bottleneck.
How does AIOps eliminate that bottleneck?
AIOps technology uses big data and machine learning to optimize IT operations.
An AIOps solution can help IT operations in many ways. It can
process a huge volume of event and performance metrics data and
distill it down to a few actionable alerts. The ServiceNow solution,
for example, reduces event noise by removing redundancy across the
board. Our proprietary machine-learning algorithms then further
prioritize the alerts.
The next step is to assist in root cause analysis. By providing all
relevant past data, metrics and rich insights around the prioritized
alerts, we guide the user to the causal conditions and the root cause.
The combination of these two steps eliminates the need for our
customers to work across multiple tools. That saves them time and effort.
The next question is, “How do we help IT operations fix the root
cause?” A smart solution should also be able to guide the user to
meaningful actions. ServiceNow takes this to the next level. We
provide potential remediation options that can be automatically
orchestrated. We also provide opportunities for automation across
different teams in IT.
At ServiceNow, our perspective is that with the help of AIOps, tasks
related to IT operations can be executed faster and in a more
efficient way, even though the tasks themselves do not change
materially. A good AIOps solution should lead customers to outcomes
that improve their operational metrics significantly.
How is ServiceNow different and what outcomes does it drive?
ServiceNow is unique in how we approach the problem. We don’t just
apply machine learning to real-time data. We also leverage historical
data, service context and learnings from human behavior. For example,
we know who changed configuration data and how. We know what worked
and what didn’t for past remediations. We learn from all of this and
apply it to future recommendations.
ServiceNow is able to do this because all that data lives on our
platform. When we apply AIOps to such rich data, we can drive more
actionable and meaningful outcomes. Examples include much higher noise
reduction, much faster mean time to resolution, ability to meet SLAs
at scale, significant reduction in the number of service issues or
outages, and significant reduction in P1/P2 incidents. Finally,
automation is driving significant reduction in manhours required to
manage IT operations.
Further validation comes from our customers and from our internal IT
operations. Symantec tells us that overall, they are seeing 95%
reduction in outages. TransAlta tells us that by the time their users
notice an issue, they are already working on it. They have achieved
80% reduction in outages. Here at ServiceNow, we’ve seen about a 67%
reduction in P1 and P2 incidents from deploying our AIOps products in-house.
Does AIOps automate service issue remediation?
Yes. The way we think about remediation is, “What are the tasks that
require human intervention and what steps can be automated?” Our
system provides recommendations on what sequence of steps should be
taken and in what order. We also give users the choice to orchestrate
remediation automatically. We make all this possible by learning from
past human actions.
We constantly aim to reduce manual effort. Our software learns from
every IT operations action, and feeds that learning into our models.
Where do you see AIOps headed in coming years?
AIOps technology can make IT operations system smart and
self-running. These intelligent systems will be able to identify and
prevent issues before they occur. With AIOps, customers should be able
to update their business services frequently and scale them globally
while continuing to meet demanding SLAs.
AIOps is also opening doors to automation beyond IT operations. Many
customers adopt the Now Platform because it uses a common data model
that feeds multiple services and also enables integration with
3rd party services. This allows customers to automate and
build workflows across IT and the broader enterprise.
When all IT process data and operations data come together, the
AIOps engine gets context it never had before. It learns the triggers
and symptoms of issues and combines that knowledge with necessary
steps in service management. Incident categorization, rerouting,
response and resolution all become automated and self-running.
Finally, intelligence and user experience go hand in hand. At
ServiceNow we focus on how users experience intelligent technologies.
We consider intuitive interfaces, guided flows and potential for smart
automation in everything we do. When you combine AIOps with natural
language understanding, users can move even faster and have a better
experience. The possibilities are endless here.