DTW (Digital Transformation World)
As AI components move into CSPs’ operational environments, we need to define how the technology should be applied, governed, managed and road mapped. The TM Forum AIOps Service Management collaboration team is doing just that.
DTWS: Why do CSPs need AIOps for service management?
This article covers the session entitled 'AIOps: Redesigning operations processes to exploit the full potential of AI' on November 11 which was part of TM Forum's Digital Transformation World Series (DTWS). This session falls under the Forum's AI, Data & Analytics theme. AIOps has heated up this year, especially since Gartner released its AIOps platforms market guide in December 2019. For communication service providers (CSPs), AIOps’ scope covers IT, networks and operations, and AI plays by different rules than traditional software. As AI components move into CSPs’ operational environments, we need to define how the technology should be applied, governed, managed and road mapped. The TM Forum AIOps Service Management collaboration team is doing just that. Leaders from the CSP and supplier communities sat down with Aaron Boasman-Patel, Vice President AI and Customer Experience, TM Forum, to discuss why a new AIOps service management framework is needed now.
“AI software is different from traditional software,” explained Luca Franco Varvello, ICT Senior Consultant for Huawei. “There are a lot of gaps in how to manage traditional software and how to manage AI software, so the new AIOps service management framework shall address those gaps and complement existing operations,” he said.
Zhang Ke, Senior Project Manager for China Telecom added that he sees AIOps being the long-term evolution path for operations functions like automated service assurance, meaning it will be a hot topic for the foreseeable future, not a passing phase.
Derek Chen, Assistant Vice President of Customer Service for Hong Kong Telecom (HKT), urged CSPs to consider not only the impact of AI in operations infrastructure, but also as part of the product set. “It’s AI for AI,” Chen said, “if you don’t have AI operations that support AI products you are crippled.”
Automated CSP operations are typically deterministic, predictable, and repeatable. AI, however, introduces non-deterministic software logic and intent-based systems to power end-to-end and even closed-loop automation. As result, “we have to move away from a traditional way of operating toward AI automation,” said Tayeb Ben Meriem, Coordinator of OSS Standardization at Orange.
“That means we have to break silos that exist today – for instance from fulfillment and assurance – and we need to integrate all of this into a framework,” he said, which is the AIOps service management working group’s aim,” he added.
“AIOps is part of the whole AI strategy in Hong Kong Telecom,” said Chen. He pointed out that in the past year HKT’s over-the-top (OTT) apps business grew 26% while its premium customer base grew 8%. Hence a reactive customer support process for OTT applications was not sustainable, so HKT applied AI to predict, prevent and analyze events to fix problems before customers noticed. “AIOps happens before the customer complains,” said Chen.
He added that to get started with AIOps CSPs need to look at three key areas: “We have hundreds of millions of users at China Telecom so it’s important for us to real-time monitor the customer experience,” stated Zhang. He said the AIOps service management framework provides “a full checklist not only in redesigning traditional operations processes but also in a way that the AI-enabled software systems are monitored, controlled, and governed.” This session falls under the Forum's AI, Data & Analytics theme which was sponsored by Cloudera