One of the necessary elements for artificial intelligence (AI) is data -- the more data available, the greater the opportunity for success with AI outcomes. What companies need for their AI initiatives is “big data” -- a large volume of both structured and unstructured data, some of which can come from UC and contact center infrastructure.
However, many companies only gather data from one site. So, curious whether more comprehensive AI data would be available from a cloud-based UC and contact center management service, I contacted Ross Williams, COO of Virsae. Using stats from November 2018, Virsae monitored 357 million calls globally for call quality, Williams said. He shared these mean opinion score (MOS) results:
- 353.3 million calls (99%) rated a 4.0 or higher MOS -- very good
- 2.9 million calls (0.8%) rated a MOS of 3.6 to 3.9 -- satisfactory
- 0.7 million calls (0.2%) rated a MOS of 3.5 -- poor
Analyzing big data from UC and contact platforms can provide hidden insights, Williams said. Companies can use the analytic results to:
- Improve customer experience and satisfaction
- Increase uptime
- Shorten the mean time to repair and resolve problems
- Increase operational efficiency
- Detect malicious and fraudulent behavior
I asked him additional questions to learn more about big data’s use and Virsae’s approach.
How do you collect the information you analyze? A small collector, which can be a turnkey appliance or application, is installed on the customer’s network to provide a secure gateway between the customer’s environment and cloud. The collector, which has little intelligence, works under the control of the cloud computing application in Microsoft Azure. Once installed, the collector reaches out to the cloud, performs a secure authentication process, and receives instruction from the Virsae service.
When you first assess the customer’s network, what do you typically discover? Understanding the customer’s configuration is key to success. The auto-discovery processes work in exactly the same way as an experienced engineer -- they use the same interfaces, and run the same commands. We use AI to analyze the returned data. This automated process runs at a pace that humans simply can’t. Within 12 hours, we discover every asset, every piece of hardware and software -- versions, serial numbers, IP addresses, MAC addresses, and physical locations.
We expand the search to locate critical touchpoints and dependencies between assets. An example of this is the humble IP endpoint, where getting the configuration right is crucial to good quality voice. From collected data, we know the hardware type, firmware version, serial number, IP address, MAC address, physical location, speed, and duplex of the Ethernet ports, V-LAN assignment, quality-of-service settings, and many more aspects.
Don’t customers already have access to this data? In the majority of cases, customers and MSPs don’t fully know what’s deployed, where it is, or how it fits together. We place this information in their hands in an easily consumed format. This ranges from Visio diagrams to federated databases and Excel spreadsheets.
How much improvement can a customer expect after using your analysis to make network changes? We always uncover issues. They fall into two categories:
- Known, and have been unresolved for a long time
- Unknown, as the customer is dependent on faults being reported by alarms or end users, or worse still, the customer’s customer
Using the additional data and actionable recommendations provided by the AI engine, we can tackle these issues, improving everything from customer experience to legal compliance. Operational people are freed up to work on higher-value work like projects. The service does most of the work.
Have you had an occasion in which the customer’s network was acceptable but later the performance deteriorated? Some have a perception that pre-deployment network verification suffices. While it’s really important for day-one operation, the reality is that pre-deployment verification provides a green light based on a snapshot in time. In the IT world, change is the only constant, and ongoing management of these environments is essential.
For example, a customer recently resolved an issue whereby it had enjoyed great network performance right up until it introduced video. Prior to video, there was little or no contention for bandwidth or network resources, and in fact QoS was not fully set up. But when another bandwidth hungry real-time application was introduced, quality went up in smoke.
Our service quickly pointed out that the issue was being caused by a single Layer 3 device not respecting the QoS model. The customer engaged suitable expertise, and normal service resumed.
What does Virsae do for customers? Virsae Service Management keeps UC and contact center systems performing well. UC and contact center systems carry complex real-time interactions; many components have to work in unison to produce a great customer experience. Traditional IT tools are limited in these environments. IT teams struggle to find and fix problems. When issues occur, their customers often chose companies that were easier to contact.
What are your customer demographics? The larger and more complex the environment, the better. Our service fits companies of any size that have adopted a culture of top quality and service excellence. It provides benefits for MSPs with operational efficiencies and differentiation over their rivals.
