Tech Insights

Two uses cases: How AIOps is solving problems for global customer service operations

Learn how IT operations teams for global customer service operations in high-tech, finance, and other industries are turning to AI-based tools and approaches to solve operations problems that are both human and machine in nature.

AI Operations-HPE Pointnext.png

These days, the rapid growth of cloud and hybrid infrastructure adoption is fueling demand for AIOps. Pre-pandemic, industry studies were predicting in fact that the AI operations (or AIOps) platform market size would grow from USD 1.73 billion in 2017 to USD 11.02 billion by 2023, at a compound annual growth rate of 34 percent.*

 One IT operations area where the adoption of artificial intelligence (AI) is taking off is in customer service and operations. Today at HPE Pointnext Services, we are working with clients to transform customer support with AI in ways that were not possible even a sort time ago.

To that point, here are two current use cases that demonstrate how introducing AI solutions into the equation can improve customer interactions, positively impact employee retention and satisfaction, and save valuable time and technology costs.

These use cases are based on experiences in two clients’ customer supports centers. One focuses on tackling challenges for employees working in a call center within the global services division of a major technology company. The other deals with server hardware issues running in a large data center used by an internal IT operations team within an international financial services firm.

How AIOps helps tackle human challenges for this global services organization

It’s no secret the employees working at call centers don’t tend to stay in their positions long, often leaving after eight months or less. The big challenge that created for our client, and one likely experienced by every customer support organization, is that this high employee churn means that no knowledge base or baseline competency level gets developed.

With each and every customer call, the customer service representative was on the front lines responding to that call—and needed to make quick decisions on how to best solve problems that can range from simple to very complex. That service rep generally had three options: One, solve the problem on the phone (costing our client around $10-to-$15 per call). Two, send a replacement part to the customer site to be installed by the customer (costing our client an average of $100-to-$200 each time). Or three, send an engineer to fix the problem onsite (costing our client as much as $600 per visit).

HPE Pointnext AI consultants worked with this client to design a tool that would predict the best (and most cost-efficient) way to solve the problem presented with each call using an AI-powered knowledge base. This approach collected and propagated data from every call center response, building a data-filled knowledge base that would continue to grow and evolve even as employees came and went. New employees joining the call center would be able to tap into that by typing a priority keyword that would take them quickly to the information they needed to make faster, cost-effective decisions.

The overarching goal was to predict if and when the more expensive call response option (sending out the engineer) was needed and work to avoid that when possible (vs. having it be a default response by more inexperienced call center employees).

We worked with the client to assess and clean the data to make sure it was in the preferred format. We then built the model for this use case. When ready, we provided the client’s team with an API to run the AI-based tool on their own.

How AIOps helps tackle machine challenges for this ITOps group within a global financial services company

In this use case, our financial services client had a hardware problem with its internal bank servers: as many as 30,000 servers out of 150,000 servers were crashing up to four times a day. The data center infrastructure here  was inherently complex and the organization’s own customer support team did not always have a clear idea of how many applications were running (and they had no idea what was causing the frequent server crashes).

At this point, the financial company’s operations team was unsuccessfully pursuing a “white box analytics” approach to this issue (which means they needed hard-to-get access to all the actual application code to be able read it and try to solve the problem. Ultimately, they ended up calling in external vendors to reboot the servers—a very expensive fix with still no clue as to the root causes of the problem. They went so far as to spend millions of dollars to double their number of servers as a way to work around unexplained downtime.

Experienced AI specialists and data scientists with HPE Pointnext came in and suggested taking a “closed box analytics” approach instead (which means observing the systems from the outside). Using AI-based analytics tools, we worked with them to pull and review all the after-the-fact system data. We analyzed large volumes of code and contrary to expectation, we discovered that the servers were crashing when running normal workloads—as opposed to crashing during peak demand.

Ultimately, we narrowed down variables to pinpoint the top five reasons that might be triggering the hardware crashes. We then handed off back to the client so they could continue running closed box analytics to gain more meaningful understanding of their own server systems.

Employing a proven step-by-step methodology to speed AI design and deployment

 In the use cases just mentioned, HPE AI and data experts on the Pointnext team used our three-point methodology to guide our clients through their specific AIOps journeys focused on operations problem-solving and improvements.

  1. Explore—We work with clients to understand the outcomes and challenges AI brings. We ground teams on common AI terminology, fostering shared understanding and selecting the best use cases. The goal is to clearly align technology with the business, so the initiative benefits from having the business buy in early on.
  2. Experience—We identify the data sources that will be required for the use case and create a high-level roadmap for use case implementation. This is followed by a proof of value (POV) as to how the solution would be deployed into a production environment. This POV is tested and the outcome is validated.
  3. Evolve—We then work with clients to evolve and scale the AI solution. Leveraging HPE’s optimized infrastructure that spans from AI edge to cloud coupled with HPE GreenLake pay-per-use consumption models makes this a much easier part of the complete journey.

Learn more

HPE Discover Virtual Experience session: Build an AI-Enabled ITOps Experience With HPE Pointnext and InfoSight

HPE Pointnext brief: HPE Artificial Intelligence Transformation Workshop

* AIOps Platform Market by Component, Service (Implementation, Consulting, and Managed Services), Application (Real-time Analytics, Infrastructure Management, and Application Performance Management), Vertical, and Region - Global Forecast to 2023; MarketsandMarkets; July 2018

Senthil Gandhi
Hewlett Packard Enterprise

About the Author


Senthil Gandhi works with the HPE Pointnext Worldwide AI & Data Division as a principal scientist leading AI efforts across the globe. He has a couple of decades of experience helping major corporations as well as startups with their AI efforts. He was nominated for Autodesk's Innovator of the Year award for two years in a row and won the award for the year 2016.

HPE Webinars
Find out about the latest live broadcasts and on-demand webinars
Read more
Online Expert Days
Visit this forum and get the schedules for online Expert Days where you can talk to HPE product experts, R&D and support team members and get answers...
Read more
View all