Are your AIOps investments effective?

AIOps has historically leveraged data analytics and automation capabilities for IT operational efficiency and associated customer value. More recently, thanks to advances in big data and machine learning, AIOps has leveraged higher levels of intelligence and insights to enable autonomous networks and digital transformation. Telcos and industry bodies such as TM Forum now play a prominent role in advancing AIOps for efficiency, value generation and sustainability. This necessitates a move away from tactical AIOps of the past, to a more structured approach, that is part of an organization’s digital transformation strategy.

A structured AIOps approach ensures balance between the needs of an organization, its customers and shareholders, and society in general. This multi-faceted approach needs to address the multiple dimensions of digital transformation - strategy, outcomes, technology, organization, data and governance – in the form of the following key considerations. These should be considered in totality to ensure success of AIOps investments.

AIOps key considerations

Bottom-up innovation

AIOps has historically taken a bottom-up innovation approach involving a patch work of capabilities that are added to legacy and new ways of working. Going forward this bottom-up approach needs to be enhanced to include the larger ecosystem whereby partners and customers have a space to innovate and reap the benefits of innovation. Standards such as TM Forum’s open digital architecture (ODA) help create a common framework for such alignment.

Democratizing data

Democratizing data ensures data is accessible and understandable not only to (few) experts but to many users, thus increasing operational scalability and success of AIOps use cases. Some of the considerations for democratizing data include access to quality data for easy abstraction, easy self-service capabilities, collaboration and knowledge sharing. For example, network fault trend analysis needs to consume data from network, tickets and other performance reports. Lack of readily available data for easy abstraction can delay insights and lead to expensive delays e.g. an outage.

Distributed deployment

Similar to cloud deployment consideration, AIOps use case deployments need to find a right balance of deploying between needs for on-prem, edge, regional, central, public/private deployments. This can be defined considering business needs, regulatory constraints, cost effectiveness, scalability and reliability of AIOps use cases. For example, AI enabled virtual assistants may be more cost effective, efficient and scalable in a central or public cloud, rather than in an on-prem deployment.

ROI Outcomes

The traditional return on investment (ROI) outcomes have focused on returns from reduced costs and/or immediate revenue gain. AIOps outcomes need to be balanced not only against cost reduction and revenue, but also on improved perceived value by customers, employees and as a differentiated service. For example, a sentiment analysis use case for customer care operations, helps assess customer concerns and prioritize them, which ultimately leads to customer satisfaction and brand value.

Operations Readiness

At times AIOps uses cases that appear sound in development and test, end up being unsuccessful on field implementation. The reasons could be several: issues with real data, lack of customer/user/operator training, inadequate consideration of potential benefits, conflicting/competing use cases, lack of alignment with technological developments (e.g. “I can do it easier and faster with ChatGPT”). A structured approach to operational readiness of AIOps use cases that includes testing, governance, training, oversight and support is essential to maximize their likelihood of success. It is also essential for such a program to ensure that various AIOps use cases support and enhance each other rather than impede one another. A readiness program should also ensure that the capability level of the targeted users and maturity level of competing technology are considered when rolling out these use cases.

Balanced AIOps governance

Similar to AI in general, AIOps should ensure a balance between the need for control vs. empowerment. While the promise of AIOps can be achieved only by empowered collaboration, the risks require it to also be adequately controlled. Typically, the focus should on centralized controls to ensure use case alignment, security, intellectual property protection, regulatory compliance, and AI responsibility while leaving units to freely collaborate and innovate within these guardrails. For example, LLM Models should have the needed guardrails like IPR protection, but should be allowed to be exploited by units for various operational use cases like co-pilots, service orchestration etc.

Using a future back approach, and by leveraging strong technology expertise and techno-economic modelling capabilities in both telecom and IT, Bell Labs Consulting helps telcos get the most out of AIOps by:

1. Defining a future state of AIOps operating model. Such a model should consider the key capabilities of technology, organization, data and service innovation, as well as congruence across these capabilities.
2. Assessing current state of AIOps operating model and benchmark against industry trends.
3. Identifying gaps to reach the future state, (i.e. relevant use cases, investments, technology architecture, etc.) ideally with associated cost and benefit analysis.
4. Defining and implementing a roadmap that leverages quick wins in the short term, while maintaining focus on the overall evolution in the mid and long terms.
5. Establishing a transformation office to support, refine and course correct use cases during implementation.

Are you are struggling with your AIOps investments or want to kick start your AIOps journey?

Learn more on how we can help you by contacting us at info.query@bell-labs-consulting.com and learn about Bell Labs Consulting at https://www.bell-labs.com/consulting/