Thursday, 12 December 2024

Harnessing AI in Telcos: Data, Opportunities, and the Future

At the Telecoms Europe virtual event, How AI is Transforming Telecoms (27 Nov 2024), Daniel Dobos, Research Director at Swisscom, presented an engaging talk titled "Have telcos got the right data in the right places to harness AI?" As Research Director, Daniel manages Swisscom’s collaborations with universities, applied sciences institutions, and research organizations. Alongside employees from various Swisscom business units, his team develops solutions leveraging cutting-edge research and technology for the benefit of customers. His prior experience includes leading AI research projects at CERN and the United Nations.

In his presentation, Daniel addressed a critical question: Do telcos have the right data infrastructure to harness AI effectively? This topic is especially pertinent as AI—especially generative AI—rapidly reshapes the telecommunications industry. He explored this question through the lenses of academic research, public sector regulation, and private sector application.

The Tsunami of AI: Opportunities and Risks

Daniel began by comparing the AI revolution to The Great Wave off Kanagawa, the iconic Japanese woodblock print. He emphasized AI's transformative potential, particularly large language models (LLMs), and urged telcos to adapt swiftly to this evolving landscape. He posed a thought-provoking question: "Are we the last generation to ask whether telcos have the right data to harness AI?" Daniel warned that telcos failing to adapt risk irrelevance or obsolescence.

Swisscom's AI Journey: From Research to Application

Daniel recounted Swisscom’s AI journey, beginning in 2016 with a partnership with EPFL to establish a digital lab focused on applied AI research for telcos. This collaboration leverages Swisscom’s extensive data resources and partnerships with leading institutions like EPFL and EDAP. Key AI techniques—natural language processing, automated speech recognition, and distributed machine learning—are deployed to enhance customer experience, network operations, and workforce management.

Real-World Applications: Optimizing Networks and Customer Care

Daniel shared three practical examples of Swisscom's AI implementation:

  1. Network Optimization: AI-driven analysis of network data to optimize mobile antenna deployment, improving both coverage and energy efficiency.
  2. Event Sequence Detection: Identifying patterns in massive data streams to predict network issues, optimize maintenance, and bolster cybersecurity.
  3. Generative AI for Customer Care: Deploying LLM-based chatbots to automate customer interactions, enhancing self-service capabilities and boosting customer satisfaction.

Knowledge Graphs: Decoding Complex Infrastructures

Daniel highlighted the role of knowledge graphs in visualizing Swisscom’s intricate infrastructure. These graphs map out infrastructure components and their relationships, enabling better human comprehension of network structures and anomalies. This approach enhances maintenance efficiency and risk detection.

Current Research: Pushing AI Boundaries

He also shared insights into Swisscom’s ongoing research in areas like infrastructure optimization, energy savings, and secure data analytics. A key innovation is the use of synthetic data, which enables collaboration between researchers and telcos without compromising privacy. Other projects include graph anomaly detection and advanced energy-saving strategies.

Key Takeaways: Data is the Key to AI Success

Daniel concluded by emphasizing that while Swisscom is well-positioned to leverage AI, many telcos are not. He underscored the importance of making data accessible and ready for AI algorithms. He urged telcos to prepare their data now, as the window of opportunity is narrowing. Comparing the current AI landscape to "teenage love," he advised experimenting with different AI models without committing prematurely. By ensuring data readiness, telcos can fully capitalize on AI’s transformative potential.

The video of his talk is embedded below:

Note: This blog post was created with assistance from ScreenApp for generating the video summary.

Related Posts

No comments:

Post a Comment