Cisco wanted to scale its digital help engineer that assists its technical help groups all over the world. By leveraging its personal Splunk expertise, Cisco was in a position to scale the AI assistant to help greater than 1M instances and unencumber engineers to focus on extra complicated instances, making a 93+% buyer satisfaction score, and making certain the essential help continues operating within the face of any disruption.
If you happen to’ve ever opened a help case with Cisco, it’s doubtless that the Technical Help Middle (TAC) got here to your rescue. This around-the-clock, award-winning technical help crew providers on-line and over-the-phone help to all of Cisco’s prospects, companions, and distributors. Actually, it handles 1.5 million instances all over the world yearly.
Fast, correct, and constant help is essential to making certain the shopper satisfaction that helps us preserve our excessive requirements and develop our enterprise. Nevertheless, major occasions like essential vulnerabilities or outages can trigger spikes within the quantity of instances that slow response occasions and shortly swamp our TAC groups, affecting buyer satisfaction consequently. we’ll dive into the AI-powered help assistant that assists to ease this difficulty, in addition to how we used our personal Splunk expertise to scale its caseload and enhance our digital resilience.
Constructing an AI Assistant for Help
crew of elite TAC engineers with a ardour for innovation set out to construct an answer that would speed up difficulty decision occasions by increaseing an engineers’ means to detect and remedy buyer issues. the was created — it’s greater than an AI bot and fewer than a human, designed to work alongside the human engineer.
Fig. 1: All instances are analyzed and directed to the AI Assistant for Help or the human engineer based mostly on which is most acceptable for decision.
By straight plugging into the case routing system to investigate each case that is available in, the AI Assistant for Help evaluates which of them it may well simply assist remedy, together with license transactions and procedural issues, and responds on to prospects of their most popular language.
With such nice success, we set our eyes on much more help for our engineers and prospects. Whereas the AI Assistant for Help was initially conceived to assist with the high-volume occasions that create a major inflow of instances, it shortly expanded to incorporate extra day-to-day buyer points, serving to to scale back response occasions and imply time to decision whereas constantly sustaining a 93+% buyer satisfaction rating.
Nevertheless, as the usage of the AI Assistant grew, so did the complexity and quantity of instances it dealt with. An answer that when dealt with 10-12 instances a day shortly ballooned into tons of, outgrowing the methodology initially in place for monitoring workflows and sifting by means of log information.
Initially, we created a strategy generally known as “breadcrumbs” that we tracked by means of a WebEx area. These “breadcrumbs,” or actions taken by the AI Assistant for Help throughout a case from finish to finish, have been dropped into the area so we might manually return by means of the workflows to troubleshoot. When our assistant was solely taking a small quantity instances a day, this was all we wanted.
The issue was it couldn’t scale. Because the assistant started taking up tons of of instances a day, we outgrew the size at which our “breadcrumbs” methodology was efficient, and it was not possible for us to handle as people.
Figuring out the place, when, and why one thing went flawed had change into a time-consuming problem for the groups working the assistant. We shortly realized we wanted to:
- Implement a brand new methodology that would scale with our operations
- Discover a answer that would offer traceability and guarantee compliance
Scaling the AI Assistant for Help with Splunk
We determined to construct out a logging methodology utilizing Splunk, the place we might drop log messages into the platform and construct a dashboard with case quantity as an index. As a substitute of manually sifting by means of our “breadcrumbs,” we might instantaneously find the instances and workflows we wanted to hint the actions taken by the assistant. The troubleshooting that may have taken us hours with our unique methodology may very well be achieved in seconds with Splunk.
The Splunk platform affords a strong and scalable answer for monitoring and logging that allows the capabilities required for extra environment friendly information administration and troubleshooting. Its means to ingest massive volumes of knowledge at excessive charges was essential for our operations. As an trade chief in case search indexing and information ingestion, Splunk might simply handle the elevated information circulation and operational calls for that our earlier methodology couldn’t.
Tangible advantages of Splunk
Splunk unlocked a stage of resiliency for our AI Assistant for Help that positively impacted our engineers, prospects, and enterprise.
Fig. 2: The Splunk dashboard affords clear visibility into features to make sure optimized efficiency and stability.
With Splunk, we now have:
- Scalability and effectivity: Splunk screens the assistant’s actions to make sure it’s working appropriately and offers the flexibility for TAC engineers to watch and troubleshoot workflows, permitting the assistant to effectively scale. The AI Assistant for Help has efficiently labored on over a million instances up to now.
- Enhanced visibility: With dashboards that permit for fast entry to case histories and workflow logs of our assistant, the TAC engineers overseeing the processes save time on case evaluations to ship quicker than ever buyer help.
- Optimized processes with real-time metrics: The visibility into useful resource allocation permits us to optimize our enterprise processes and workflows, in addition to reveal the worth of our answer with real-time metrics.
- Proactive monitoring: Splunk ensures all APIs are totally functioning and screens logs to alert us of potential points that would affect our AI Assistant’s means to function, permitting for fast remediation earlier than buyer expertise is impacted.
- Greater worker and buyer satisfaction: Engineers are outfitted to deal with larger caseloads and effectively reprioritize efforts, decreasing burnout whereas optimizing buyer expertise.
- Diminished complexity: The dashboards have a easy interface, making it a lot simpler to coach and onboard new staff. The convenience of use additionally serves to enhance the capabilities of the people working our AI Assistant by enhancing their accuracy and effectivity.
By offering a scalable and traceable answer that helps us keep compliant, Splunk has enabled us to keep up our dedication to distinctive customer support by means of our AI Assistant for Help.
Further Assets:
PS: Attending Cisco Dwell in San Diego this June?
You’ll have a particular alternative to speak stay with Cisco IT consultants to dive into these success tales and different deployments! Look for Cisco on Cisco in every of the showcases and remember to search Cisco on Cisco within the session catalog to add our classes to your schedule!
Share: