1xpert - Fotolia

Guest Post

How AIOps investment improves customer service, boosts ROI

Before reaping the benefits of AIOPs -- such as improved customer service and easier collection of customer data -- companies need to determine if it is a good fit.

In most enterprise models, investment in infrastructure follows a simple rule: spend to add value. In that view, the IT operations investment is relatively low, and its average annual spend as percentage of revenue was just over eight percent, according to the "Flexera 2020 State of Tech Spend Report."

That kind of budget might take care of low-hanging fruit such as anomaly surveillance, security and network maintenance, solving events and customer service queries. However, with a longer-range investment, IT operations could deliver better customer service and data-driven products.

Proliferation of technology

Despite the proliferation of technology in everyday life, most customer service interaction is still handled through a manual process of customers raising tickets which are then resolved within IT operations. Integrating AIOps -- a combination of artificial intelligence (AI) and operations -- offers a more efficient way to manage the volume of tickets, allowing quick solution to customer issues -- customer view from outside in -- and a means to gain insight into customer needs -- enterprise view from within. For example, some enterprises have invested in bots -- a byproduct of AIOps implementation -- which allow operations teams, freed from large numbers of repetitive customer tickets, to focus on other initiatives such as optimizing customer experience through predictive analysis.

With AIOps, operations can surpass that reactive management of customer issues to proactive improvement of customer experience and further forward to product innovation through data analytics. This is achieved through automation via AIOps, which can speedily manage customer tickets, security scans and gather data on customer needs at the same time.

AI has negative connotations for many, including the belief that "robots" will replace humans. AIOps is not intended to replace personnel, but instead should be looked at as an opportunity for IT operations to reskill or upskill its team members and allow them to focus on what comes next through innovation. AIOps can help collect and aggregate large and ever-increasing volumes of operations data generated by multiple IT infrastructure applications, and customer usage patterns that would typically take IT operations teams thousands of hours to manage. That is too much data to manage effectively without some sort of AI intervention.

AIOps aids in the diagnosis of root causes of issues and can rapidly respond and remediate with self-healing solutions without human intervention.

Traditional IT tools can't keep up with the volume and don't allow for scalability based on the demand, due to the lack of insights to correlate data across different but independent systems and environments. Real-time insights become next to impossible in traditional IT operations.

The introduction of AIOps closes the visibility gap in IT infrastructure, across dependencies, and provides a view on the usage patterns of the customer. AIOps helps in quick remediation of slow-downs or outages and can alert IT operations to problems and recommend solutions. At the same time, AIOps can continuously scan for security threats or run tests to look for vulnerabilities.

Investing in technology

AI has become integral to tracking customer preferences and trends -- take Amazon, for example. The data is used to suggest products to shoppers but, if tracked, the data can also serve as a guide for retailers and fulfillment centers that interest in certain products has spiked, and to stock accordingly.

IT -- if expanded to proactive processes -- can be thought of as a parallel support to launching new products or services. There are a couple reasons for this:

  • High customer expectations. Customers have come to expect speed and reliability with online apps and websites. It can damage a brand and sales if functionality falls below expectations.
  • Frequently changing devices. Platforms and devices continue to change. The devices which customers use to access sites and apps are continuously being updated with new OS versions and models.

IT operations needs to be able to react to changes, if not anticipate them. From an IT perspective, this means agility through reduced complexity, and a better view of IT infrastructure itself. Proactive review of usage trends can help an IT team foresee risk and improve the platform before it becomes compromised. These activities encompass AIOps if these tasks are automated using machine learning and patterns that have been developed based on customer activity data.

Transition to a new model

Transitioning from a traditional IT operations model to AIOps is not a one-size-fits all process.

Because AI and machine learning can process large amounts of information, AIOps can mine data on customer activity and inform strategy. Broadly, AIOps can assist with evaluating capacity, monitoring resource and storage management and detecting system threats. Transitioning from a traditional IT operations model to AIOps is not a one-size-fits-all process. It requires a long-range investment, unlikely to pay off for three to five years. And that is if AIOps is a good fit for the organization.

There are some important questions to consider when evaluating if AIOps is the right fit for an organization, such as:

  • Does the use case call for AIOps? If there is not a compelling reason for the integration, it is time to head back to the drawing board. The use case needs to be fit for purpose and to have a long view on the end goal about three to five years down the road.
  • Is the AIOps strategy adaptable? The strategy must be able to change as integration moves forward. An IT operations team will likely pivot and adjust as testing and experiments bring to light new problems or possibilities.

Choosing a use case that is a mismatch for AIOps is a common misstep, and a good way to burn through added budget with little return. The best way to avoid this blunder is though data analysis. An organization with years of quality customer activity data will be well-positioned to understand how AIOps can best be used. Many firms have not been collecting data long enough, or the data may be "noisy" with confounding information.

Once the use case and health of data have been verified, some best practices can help keep an AIOps rollout on track:

  • tag and catalog data;
  • rank both services and dependencies; and
  • seek expertise for portions of the process, often with data analyses.

Many organizations do not currently have the capability for robust data analysis. Rather than attempting to build skills from the ground up, it may be faster, cheaper and wiser to find an expert.

Reap the rewards

Understanding the elements involved, a company needs to evaluate if AIOps will add value. Evolution is inevitable, and IT operations is no exception. Customers are utilizing an ever-broadening range of devices and platforms which also continue to evolve. Organizations need to be ready to meet the needs of customers. AIOps can prepare an enterprise for continued changes in technology as well as lead to a better position in the market.

With increased and intelligent spending in AIOps, an organization is prepared to provide quality customer service and added value to business objectives. Streamlined AIOps provides an advantage against competitors, identifies areas of business growth and turns operations into a value stream rather than a cost stream.

About the author
Srinivasa Wudaru is the vice president of information technology for a major financial services company. With an MBA in product development and innovation and a Bachelor of Technology in computer science, he has twelve years of experience in developing AIOps programs to support his team and organization. For more information, please email [email protected].

Dig Deeper on Customer service and contact center

Content Management
Unified Communications
Data Management
Enterprise AI