5 Ways to Solve IT Infrastructure Errors in The Call Center
By Kelly Koelliker
Posted in Customer Engagement
First: Call centers are spending far too much time and money trying to make sure their communications infrastructure is running properly.
Next: For most large companies, your call center is absolutely critical to the success of your business. But the success of these exceedingly important calls is completely dependent on the infrastructure supporting them -- often an incredibly complex, tangled web of systems and vendors.
Here's the Challenge:
Because of complexity mentioned above, companies struggle to find the resources required to keep the systems running, ensuring calls are correctly received, routed and recorded.
This leads us to the Big question:
How can you be sure your call center is running at peak performance without blowing your budget on IT resources? What is the right testing or monitoring approach? Read on to find out.
Before getting into the top tips of how to monitor and test your call center infrastructure, it's important to understand what these terms mean and how they are different from each other.
First: Testing is an active process. Through scripts and applications, you can stress the system, performing a large number of transactions within a specified period of time.
Next: Testing will discover issues and errors during the test, but you can't test at all times. So during times that a test is not running, new issues may crop up.
Dealing with the Element Management Factor
Here's the thing:
Another challenge with testing is the incomplete nature of a type of testing called element management. In element management, each individual component, or element, is tested. Without a complete view, element management can yield inaccurate results.
Here's an example:
First: in a call recording environment, element testing of the recorder may show that everything is functioning properly.
Second: When the entire service is viewed as a whole, calls are not actually being recorded.
Finally: The recorder may show success because the true issue is elsewhere in the environment.
This brings us to: Comprehensive Service Testing
Comprehensive service testing can be extremely informative. A test will pass or fail, depending on what outcome is expected.
The good news:
Even if the failure is due to a situation you hadn't even though of, the test will find it. When you create a test, the script assumes the role of a human, logging into a service and attempting to use it.
Something to keep in mind:
A test doesn't care what systems are running on the back end -- applications could be in the cloud, hosted by a third party, or run locally. These components could otherwise be inaccessible to test individually. As such, the test will find any component of the system that results in a failure. The trouble, however, is that the test is often unaware of the ultimate source of failure. The test knows that it has failed, but it may not know why.
The Takeaway is this:
You must still dig back into each component to find the root cause.
First: Monitoring takes a different approach than testing.
How so? Monitoring is passive. It looks for alerts and anomalies while watching the system operate under normal conditions.
Second: While monitoring can be less disruptive, it is also not proactive.
Next: With monitoring, you won't see a problem until it actually happens.
First: With monitoring you can only watch for conditions that you know about.
Second: With monitoring, the application will have a pre-defined set of alerts that are created to fire under certain conditions.
Third: If a failure occurs that is not one of the known conditions, the monitor may not pick it up.
On the other hand, when an alert does fire, the administrator will know exactly what issue has caused it. As opposed to testing, monitoring alerts are very specific.
The Bottom line:
Monitoring has become challenging in the age of modern computing.
Next: With complex environments including components hosted in the cloud or by third parties, administrators may not even have access to monitor specific components.
Second: A single alert from a monitor may not be catastrophic. The administrator will need to understand the alerts to know if a specific condition may only cause a delay on a certain device but may be catastrophic if the condition occurs across multiple devices.
Clearly, both testing and monitoring have pros and cons. As we'll see, both elements are critical to the long term success of your telephony infrastructure.
First: In today's world, organizations need to monitor an interconnected web of hardware and applications across multiple tiers, from multiple vendors.
Next: There are literally thousands of faults to watch for. The simple assumptions of element management and element-based alerts are no longer sufficient.
Finally: Even if there are no alerts, services may not be functioning optimally. Likewise, if you see a fault, services may not necessarily be compromised.
Here's the thing: To know where to go next, we first need to look back a bit.
The first "bug" found on a computer system was an actual bug -- a moth that had flown into a server and shorted the circuit. Ever since then, people have been trying to put processes into place to identify and catch bugs.
Who remembers SNMP?
Simple Network Management Protocol (SNMP) was one of the earliest attempts at a unified monitoring system.
What did it do? Good question.
Basically . . . This was a vendor-agnostic mechanism where various applications could send alerts in a unified way.
Next: Many monitoring programs developed since then followed much of the same process. Initial monitoring programs focused on "element management," whereby the monitor looked at a single device and faults that could occur on that device.
Finally: This approach worked well in a pre-Internet era when most devices operated as independent silos.
Organizations have tried several approaches to maintain the availability and stability of their IT infrastructure. Unfortunately, many of these approaches require an extreme investment of resources for very little benefit. For instance:
Walking the Floor: In certain industries, compliance regulations dictate that each and every call across every phone must be recorded. Because of the lack of holistic monitoring and visibility across all devices, administrators resort to literally walking the floor of the call center, testing device and audio quality line by line.
Diagnosing without Data: When testing reveals a problem, administrators may only know the result -- certain calls aren't being recorded, or perhaps certain terminals aren't receiving calls. Unfortunately, the IT team lacks any further information about what may be causing the problem. Instead, they waste precious resources trying to examine each component of the system, each requiring specialized skills and expertise.
Out-of-date Tests: Sometimes, the IT department attempts to write detailed test scripts to ensure the health of the environment. But in a complex organization, things are constantly in a state of change. Users come and leave the company, change roles, and get new devices. Software is upgraded and configurations are changed. Test scripts are almost immediately out of date and more time is wasted.
Load Testing: Traditional load testing simulates a mass amount of transactions to stress the system. The trouble is that a load test doesn't actually simulate reality. In a real scenario, there are lots of different types of interactions across your networks and devices, not just the same transaction again and again.
Here's the thing: We've talked a lot about things that don't work.
But what should you actually do? Is there a way to achieve holistic monitoring of your telephony infrastructure without so much effort? There is.
First: Both testing and monitoring are helpful, but both have shortcomings.
Second: The right approach is to connect the two approaches, and power your processes with intelligent automation.
With Verint Automated Verification, automated end-to-end monitoring checks for alerts across all of your applications: databases, your ACD, recorder and other software, and the handsets themselves.
What it means for you:
First: If an alert is triggered, an auto-discovery process is able to determine what portion of the environment needs to be tested.
Next: It takes parameters from the auto-discovery to simulate the precise conditions that created the alert.
With a holistic solution like this, organizations have a complete view of their infrastructure as well as in depth analysis as needed.
First: Your IT staff are better able to find and resolve any issues that may compromise their telephony services.
Second: The result is a lower mean time to resolution, reduced costs, greater operational efficiency, and the compliance adherence your business requires.
First: Early testing and monitoring focused on individual devices.
Next: From there, tests evolved to analyze platforms, which are comprised of hardware, operating systems, and software. After platforms came service monitoring, which in turn evolved to analyzing transactions themselves.
This evolution is important because in many cases the IT staff is no longer concerned with individual devices. In fact, many companies offer a "bring your own device" policy that eliminates the device from the equation entirely.
What comes next? Moving from devices, platforms, services and transactions.
At the top of this maturity curve is people.
First: People are fundamentally what an organization cares about.
Second: People are also some of the most difficult elements to monitor and track.
Finally: In a large organization, people are constantly joining and leaving the organization, changing roles or changing locations.
If you're simply looking at individual transactions, you may never notice when things go wrong in this change.
First: You need an automated solution that sees all of the calls across all of the people.
Next: You need an automated solution that knows how each interaction should play out.
Finally: You need an automated solution that alerts you when the expected behavior does not result.
Here's the bottom line. You don't have to suffer with your complex telephony infrastructure. You don't have to worry, guess or hope that your calls are being recorded. With Verint Automated Verification, you can be sure.
Want to stay up to date with all the latest insights?
Subscribe to our weekly or monthly digests of all the latest insights and articles from from Verint