AIOps Evolution Podcast | Season 2
Interview with Bhaskar Krishnamsetty, Chief Product Officer & Tejo Prayaga, Sr. Director of Product Management and Marketing from CloudFabrix
The dialogue around AIOps integrations has shifted drastically just in the last two years. Previously, vendor conversations revolved around explaining the benefits and value of AIOps to prospective investors. Now, people believe the story, but they want to know how to make the AIOps story a reality. The prevailing question now is: ‘How do we get to frictionless AIOps?’ Bhaskar Krishnamsetty and Tejo Prayaga from CloudFabrix joined Sean on AIOps Evolution to dissect these shifts in the AIOps dialogue and what they mean down the road.
TL; DR
How the AIOps dialogue has shifted
Compared to about 18 to 24 months ago, the current state of AIOps has improved dramatically. Companies that used to pitch AIOps platforms or solutions would highlight the benefits of AIOps integrations, but now that dialogue has shifted.
Thanks to the efforts, research, and funding of organizations like Gartner, most companies know the value and benefits of AIOps. Now, the conversation revolves around how to get there. More importantly, how do we achieve frictionless AIOps maturity?
Frictionless AIOps Adoption
Bhaskar emphasizes one way to ‘measure’ frictionless AIOps adoption is the ease of implementation. There are two categories: technically frictionless and culturally frictionless. Here’s how they each contribute to a smooth AIOps adoption:
Technically frictionless
Technically frictionless refers to the time to value, data integrations, and data plumbing. All of this is part of the AIOps implementation process. It can be very time consuming and cumbersome for novice AIOps adopters. This is where vendors can step up by providing automation and technical guidance.
Culturally Frictionless
On the other hand, as with any change, AIOps is a big shift in the culture and processes of a working environment. Getting buy-in internally can be an uphill battle not only because it’s something “different”, but it’s work to train on a new process or technology. So, it takes a lot of people and training to get dedicated groups that know the AIOps platform or solution to make the transition culturally frictionless.
Bhaskar says, “Sometimes it’s a top down initiator starting with CXO and all, which is usually good because the rest of the groups align, but sometimes it is ITOps or another operations team. Sometimes it is, you know, DevOps teams. So, there is no clear cut, defined buying center at all [for AIOps].”
Bottomline: No matter where an AIOps implementation is coming from within the organization, you’ve got to have a solid strategy mapped out for a technically and culturally frictionless experience.
Taking AIOps integrations to the edge
We all know the general use cases for AIOps: anomaly detection, alert, noise detection, root cause analysis. These are key AIOps use cases, but we are seeing some evolving use cases, too, especially in DevOps and cloud native or microservices.
- Helping develop better software: As we all know, there’s a lot of logging but not much monitoring, which presents an opportunity to help during the CI/CD pipeline itself. AIOps helps the development and QA team with features, logs and operational data.
- Edge generated AIOps: Another use case that Bhaskar and Tejo have observed much is edge generated AI integrations for telecom service providers. As more technologies become 5G operable, telecom providers want to leverage AIOps to bring 5G from edge to core to cloud.
Natural Language Processing: Making AI “human understandable”
Natural language processing (NLP) is the branch of artificial intelligence or AI—concerned with giving computers the ability to understand text and spoken words in much the same way human beings can. With NLP there is a lot of unstructured and semi-structured data that is housed in logs, incidents or change requests.
This is an area where AIOps and the vendor can help end users to process and make sense of the data. While it’s a nascent AI integration, companies like CloudFabrix are developing advanced algorithms to bring NLP to the forefront of AIOps adoption.
CloudFabrix use case for NLP
Bhaskar mentioned a client in the financial sector that was using on-prem ServiceNow. Their environment was customized with specific tribal knowledge in the resolution notes and descriptions. The company was approached by ServiceNow and informed that they had to move out of the SaaS version.
This move would incur a loss of the customized fields and text they had developed over a decade. They brought CloudFabrix on to help them preserve the knowledge base, which included ten years of metadata, change requests, and incidents.
CloudFabrix applied advanced AI algorithms and NLP models to fit these customized logs. The final result? The company was able to preserve their knowledge base and seamlessly integrate to the new ServiceNow parameters. It’s more than saving them the headache of migrating or losing their system data though. Utilizing AIOps saved them in downtime, maintenance, and allowed them to keep their current NLP models. This helped with future classification, resolution, and analysis.
Vendors: Builders of AIOps integrations and the trust factor
People believe the “AIOps story”, but they want to know how to make the story a reality. The use cases are out there as well as the data to support using AIOps for greater productivity. Bhaskar, Tejo, and Sean all agree that while there is still some hesitancy, there is a trust factor around AIOps adoption.
Wherever there is an area producing a lot of data, that’s where we see the most use cases for AIOps. At this point, it’s not a question of when we adopt AIOps, but how. This is an area that vendors and AIOps platforms have to pay attention to. As the dialogue continues to shift around AIOps integrations, it’s important to highlight the business value, the best use cases for an organization, and how vendors can lead the charge to make their journey as frictionless and seamless as possible.
