The recent Digital Transformation World event saw numerous AI-related announcements that demonstrate AI's continuing maturation for CSPs.
Omdia view
Summary
Telecom IT suppliers including Amdocs, Ericsson, Netcracker, Nokia, and Salesforce (in combination with Aria Systems) made announcements around and during the recent Digital Transformation World event in Copenhagen that touched upon aspects of data management and modeling. Omdia believes this is an important moment in the maturation of AI in the telecom domain, as the market is now addressing the underlying foundations and giving CSPs usable tools to help them extract data to help them run their business more effectively.
Data models will be the next point of differentiation for telecom IT
Hoovering up the Internet, as generic large language models (LLMs) have done, isn’t of much use to a CSP’s operations team. However, accessing millions of trouble tickets or conversations with customer service agents could be. Everyone knows CSPs are sitting on a treasure trove of this type of data and that, if well managed, could provide useful insights for running a network, caring for a customer, and, ultimately, growing a business. How all that disparate data relates to each other is where the magic lies and is the secret sauce for any data model.
Suppliers’ announcements address different aspects of AI for CSPs
Amdocs
Amdocs announced its amAIz GenAI framework in June 2023 and at DTW announced Amdocs Cloud Management Platform (ACMP) using amAIz to support its delivery, quality assurance, and managed service for CSPs. Based on the amAIz framework, ACMP allows Amdocs to provide “hyper-automate[d] service development and operations, and manage and streamline a service provider’s entire IT and business operations lifecycle.” In this scenario, CSPs will reap the benefits of lower costs, higher quality and agility, and improved customer satisfaction without becoming GenAI experts themselves.
Ericsson
Ericsson announced an extension to its Ericsson Expert Analytics product, introducing an advanced troubleshooting capability powered by machine learning (ML) and artificial intelligence (AI). The platform uses “embedded system intelligence to draw selected data from a CSP’s core network and other sources for subscriber-centric analysis.”
Separately, Ericsson announced an expansion of its collaboration with Google Cloud for Cloud RAN. The Ericsson Cloud RAN solution on Google Distributed Cloud (GDC) offers integrated automation and orchestration using AI/ML. Ericsson has taken a somewhat more conservative approach to AI than some of its competitors but has targeted clear use cases with clear benefits.
Netcracker
Netcracker announced its GenAI Telco Solution the week before DTW. Built upon commercial GenAI models, Netcracker’s GenAI Telco Solution leverages its telco expertise to deliver “Knowledge Management to build, test and optimize telco-focused scenarios, such as customer care, business operations, sales and network operations.” It also supports personalized prompts for high-quality GenAI interactions. During the event, it announced expanded partnerships with both Microsoft and Google Cloud. Microsoft’s Azure Open AI Service and Google Cloud’s Vertex AI will now support Netcracker’s GenAI Telco Solution. Giving CSPs the choice of cloud providers makes good sense at this early stage of GenAI development.
Nokia
Nokia’s launch of its AVA Data Suite during DTW focuses on giving CSPs tools to help them develop their own AI/ML software applications. It “provides CSPs and data scientists standardized, pre-correlated data products packages across multivendor 4G, 5G and fixed domains to accelerate the development of AI solutions for engaged network performance and sustainability, improved subscriber experience, and new revenue streams.” It will run on Google Cloud and leverage its BigQuery data warehouse and the VertexAI platform to train AI/ML models. Getting the data ready takes an enormous amount of time, and CSPs are often overwhelmed thinking about how to get started on transforming raw data into AI/ML-ready data products. Nokia’s solution could be a great help.
Salesforce and Aria Systems
Salesforce and Aria Systems announced their partnership just ahead of DTW that has led to a “new AI-optimized concept-to-care monetization solution.” It is the result of the integration of Salesforce Customer 360 and Aria Billing Cloud and will leverage the GenAI capabilities of Salesforce’s Einstein. While Salesforce’s CRM is widely deployed across multiple verticals, it has invested significantly, via its Communications Cloud, in telco-specific capabilities that it believes give it a competitive advantage for delivering personalized user and customer experiences.
Looking ahead
With people becoming more conversant in GenAI due to their experience with ChatGPT and other LLMs, the next step is to fine-tune the foundational models to make them relevant and valuable for CSPs. Of course, any LLM requires tremendous computing power for inferencing, so while the hyperscalers are not discussed in detail in this article, they are clearly going to play a pivotal role in supporting CSPs’ GenAI ambitions.
CSPs have no shortage of companies prepared to help them along their AI journey. Those with the most robust, flexible, and usable models stand to reap the most from this once-in-a-generation opportunity.
Appendix
Further reading
The Role of AI in Customer Experience Management – Analysis (Sep 2023)
“Applying GenAI to network automation—an EnterpriseWeb proof of concept” (Aug 2023)
“GenAI in telecom operations – living up to the hype” (Jul 2023)
“CSPs are progressing on their AI journey” (Jun 2023)
AI Market Maturity Survey 2022: AI Taking Root with CSPs (Nov 2022)
Author
Roz Roseboro, Principal Analyst, Service Provider Transformation
