The conversation around AI continues to evolve and is starting to dominate discussions.
Generative AI has already had a positive impact on the telecoms industry and continues to consume attention. Meanwhile, Agentic AI is now prominent on our agendas.
How will Agentic AI play a role in telco operations – and what can it offer for service assurance?
Agentic AI has attracted considerable attention across most industries and sectors. That’s primarily because it offers two key advantages over Gen AI. While GenAI helps users discover and filter information from a mass of data, Agentic AI can enable:
Decision making
Task execution
In other words, while GenAI helps us process data at unprecedented scale and velocity, Agentic AI approaches also enable autonomous processes and extended workflows. So, as the GSMA notes:
Agentic AI empowers systems to understand intent, make autonomous decisions, and execute complex task. GSMA - Agentic AI in Telecom
Naturally, this opens up an array of opportunities – which CSPs are eager to embrace. Let’s explore some of these in this emerging landscape.
Agentic AI will be at the heart of efforts to enable network automation and to support progress towards the upper levels of the TMF autonomous network framework.
For example, any workflow that requires consideration of different data sources in order to determine the next steps can benefit from Agentic AI – including customer care, service personalization and customization, provisioning and more. In this article, however, we’ll focus on what it can offer for operations and service assurance.
Let’s start with a simple example: Alarm handling. When network alarms are generated, they need to be analyzed, and appropriate responses taken. Any given alarm may have several possible actions for resolution, so an Agentic AI tool could evaluate these and, based on the available data, take a decision and implement the required change.
However, as industry analysts Omdia have noted, many such alarm conditions can already be resolved by rules-based algorithms , using deterministic not probabilistic principles.
Regardless, Agentic AI approaches to alarm handling and other issues are sure to proliferate – and much work has already been undertaken in this domain. It is, though, important to note that, despite such promise, AI is not always the answer.