Repeat after me: AIOps for future telcos

Applying artificial intelligence to IT and network operations (AIOps) offers huge opportunities for communications service providers (CSPs). A TM Forum Catalyst proof of concept aims to show how to make the benefits real. In a recent survey by TM Forum, improving customer experience was ranked as the most significant driver for deploying AI by 77% of CSP respondents. The second most popular driver was reducing operating expenses (OpEx) through automation and closed-loop systems, followed by preventing failures and outages.

However, there are still significant challenges to achieving these aims, including inconsistent and fragmented data, immature components and support systems, lack of end-to-end service management and governance standards, trust and accountability issues and more.

AIOps for future telcos

A multi-phased TM Forum Catalyst proof of concept is looking to make these benefits real, tackle the obstacles and create a toolkit of assets, best practices and repeatable processes that can be adopted by the telecommunications industry. The AI for IT & Network Operations (AIOps) – Phase II Catalyst, which will be demonstrated at Digital Transformation World in May, follows on from project showcased at Digital Transformation Asia in Kuala Lumpur last November, which won the Outstanding New Catalyst award, highlighting its importance to the broader industry. The team explored five use cases around how to improve customer experience and efficiency by optimizing IT processes and operations. The initial phase included three Champions – China Telecom, China Unicom and China Mobile – who between them manage more than a billion customers, and three participants, BOCO Inter-Telecom, Huawei and Si-Tech. Says Luca Varvello, ICT Senior Consultant & Advisor, GTS Department, Huawei:

"The first phase aimed to show, test and demonstrate that AI technology is mature enough to be effectively infused into the operations service management processes.”

Watch this video to learn more about the first phase:

Growing the team

The second phase is extending this work, and the team has grown too. Three new champions have joined – Telecom Italia (TIM), Hong Kong Telecom (through its subsidiary PCCW Global) and Telefonica Germany – as well as two new participants, BoNC and Tech Mahindra. Catalyst champions, often CSPs, bring a business challenge that needs to be solved, and they work alongside the participants to develop rapid-fire solutions to the specific challenge over three to six months.

Massimo Banzi, Standards Manager, Telecom Italia, says one of the reasons his company joined the Catalyst is because its focus maps precisely to the future direction of the business:

“It is mandatory to use new technology as much as possible for improving the service that we offer to our customers. We need to be flexible and we need to anticipate the need of the customer. “Hence, artificial intelligence – which comes from the continuous evaluation of data collected from all of the possible sources. This means collecting data from several categories of customer and collecting data from the network to understand when there are performance issues, then tackling them by dynamically redirecting, adding and scaling network resources, etc.”

One of the key reasons Hong Kong Telecom joined the Catalyst is to deepen customer engagement through enhanced customer experience with smarter business operations, and to offer a unique integrated quad-play experience, according to the team. The company wants to be able to detect abnormal customer experience early and fix these issues before they become complaints. Through the Catalyst, they are exploring how AI algorithms can analyze collected performance data to drive predictive and corrective actions which minimize non-effective cell site visits and identify hidden problems and their causes to enable a quick fix.

Repeat after me

The new champions extend the scope of the Catalyst by bringing their individual requirements and business priorities, as well as perspectives from different markets, services and solutions. Phase two of the Catalyst will extend AIOps to additional areas of CSPs’ service operations management to show a more comprehensive, end-to-end view of how AI can benefit CSPs. The graphic below shows the service management processes that are addressed:

In addition to the deployment of new use cases based on machine learning algorithms, this phase will focus on designing AIOps frameworks, assets and repeatable practices. Says, Varvello:

“We are designing toolkits to help with the training of the AI models and the collection of data as well as integrating AI models into existing systems and architecture.”

The team will take outcomes from the AI Collaboration Project, test these through the Catalyst, and provide feedback and contributions. They will then present their findings back into the collaboration project to drive the focus of future work. This will particularly benefit the AI Program suite of tools and workstreams, including Service Management Standards, AI Data Model, AI Data Training Repository, AI Maturity Model and AI User Stories and Use Cases. As they are operational support systems (OSS), the AIOps frameworks will be aligned with TM Forum’s Open Digital Architecture (ODA), which is a vision for a more agile and flexible OSS architecture. Further, APIs developed for the frameworks will be aligned with TM Forum’s Open API data model, which defines the data structures used within TM Forum REST APIs.

How does it work?

Each member of the Catalyst team plays an important role, with champions and participants pairing up to work on specific use cases:

• Huawei is working with China Mobile, TIM and Hong Kong Telecom on cross-domain and multi-layered use cases covering multiple service management processes, from event management to customer experience management.
• Si-Tech is leading a cross-vendor team to develop the AI Ops general framework with China Mobile.
• BoCo is working with China Telecom on around predicting base station capacity.
• Bonc is collaborating with China Unicom on a use case related to faults and log analysis.
• Tech Mahindra is partnering with Telefonica Germany on AIDevOps for order fallout and incident analysis and prevention.

“Everyone is bringing different perspectives on different use cases and what is connecting us is the design of AI frameworks, platforms, best practices and reusable assets for the deployment and the operations of AI systems in CSPs’ production environments.” says Varvello.

One to watch

At Digital Transformation World next month, the team will showcase new machine learning-based use cases for specific operational processes. “Expect real demos with real data and real applications,” says Varvello. The team will also be presenting the universal standards, assets and processes they have drawn from their latest work. The AIOps Catalyst will be showcased in the new dedicated AI Zone at Digital Transformation World. The AI Zone will show how AI is driving digital transformation, with examples from leading suppliers, operators and Catalysts. Says China Telecom’s Zhang Ke:

“Our global team members are working together highly efficiently to press this project forward. We can hardly wait to see our remarkable achievements in phase 2.”

Update

Learn more by watching this video filmed at Digital Transformation World 2019: