Artificial intelligence (AI) for telcos has come a long way from a tool that simply pulls information from a database, powering clumsy digital assistants trained to answer frequently asked questions, share “canned” information or perform simple customer service tasks.
The advent of generative AI (GenAI), which uses large language models to continually learn, has allowed telecom companies to gain a 360 view of customer engagement, from first contact to buying, installation, support and billing.
GenAI offers a new path forward for revenue streams, breathing new life into an industry facing rising costs and increasing competition. With declining costs for voice services, telcos’ traditional revenue model based solely on minutes of talk time is eroding drastically, requiring them to find new services and business models.
“AI has really become pervasive throughout the entire value chain of a telecom operation,” said Silvia Candiani, Vice President of Telecommunications, Media and Gaming at Microsoft, in a recent Fierce Network interview, where she highlighted its impact on everything from customer service to network analytics and complex B2B applications.
As Gen AI gets smarter, it enables telcos to turbocharge their operations, breaking down silos, removing information bottlenecks, unlocking new revenue streams and driving never-before-seen efficiency in customer service and beyond.
Below are five examples of telecom GenAI use cases that are already driving market differentiation and bottom-line results for early adopters:
Fraud Detection and Security: Telecom fraud, from so-called “vishing” (voice phishing) that affects customers to international revenue share fraud (IRSF) and SIM box swindles that deprive operators of normal toll charges, is evolving. It’s also on the rise: In 2023, the global telecommunications industry reported $38.95 billion in fraud losses, which is a 12% increase from 2021, according to Federal Trade Commission reporting. In the first quarter of 2024, telecom fraud losses were about $250 million, up from about $180 million in the first quarter of 2021.
Earlier versions of AI and ML have enhanced telecom security by analyzing patterns in real time to identify suspicious activity in billing systems or network access based on rules-based, often-static systems. GenAI is taking it further, allowing automated, smart remediation and predictive risk mitigation.
Indeed, GenAI systems based on large language models (LLMs) can learn from historical fraud patterns and user behaviors to predict and proactively prevent future fraudulent activities. They also continuously evolve, adapting to new fraud tactics without the need for manual intervention. In terms of business outcomes, it delivers enhanced detection accuracy, almost-real-time fraud prevention, scalability and reduced false positives.
To protect telecom clients from the rising sophistication and frequency of targeted cyberattacks, Nokia has expanded its NetGuard Cybersecurity Dome offering to include a telco-centric GenAI assistant based on Microsoft Azure OpenAI Service. The assistant is trained on 5G networks and Microsoft security as well as Nokia’s telco domain expertise.
Announced in February, the GenAI assistant could reduce the time it takes to identify and resolve a threat by up to 50%. AI and automation are integrated into multiple pre-built 5G security use cases for the Core, RAN and Transport domains. These services are based on an Extended Detection and Response (XDR) capability that collects, aggregates, analyzes and correlates security data from a variety of sources, enriching it with a telco context to help security operations teams assess business risks, improve decision making, and control costs and risks.
Voice AI: Telecom innovators not only use AI to detect and safeguard their networks against fraud, but they also are infusing AI into the voice network to create new products.
Norwood Systems, an Australia-based pioneering telecom solutions company, is Integrating real-time voice AI services with its CogVoice™ solutions that enhance subscriber engagement, acquisition and retention while delivering advanced Agentic AI capabilities. Norwood leverages Azure OpenAI Service, Azure AI Translator and other services into its advanced voice solutions for telecom customers.
Another Voice AI user, Telkomsel, Indonesia’s leading digital telco company, credits the firm’s virtual assistant Veronika with 15% fewer customer escalations, resulting in a net promotor score that’s 27% above the industry average, according to a news release on how AI is transforming telecom at MWC Las Vegas 2024.
Telkomsel created Veronika with Microsoft’s Azure OpenAI Service. The highly personalized chat-based virtual assistant uses the latest information from multiple FAQs and Azure OpenAI to provide more insightful answers.
Bharat Alva, Chief Information Officer at Telkomsel, says generative AI was pivotal to creating the kind of intuitive and accurate virtual assistant needed to serve Indonesia’s diverse customers, who live on over 17,000 islands and speak more than 700 languages. “Azure OpenAI Service provided the intelligence that delivers accurate, personalized results for our customers,” he says.
The groundbreaking virtual assistant uses natural and empathic language that feels like how a human assistant would interact with clients. Since introducing Veronika, Telkomsel has observed customer self-service interactions grow from 19 percent to 45 percent. Azure OpenAI provides a scalable architecture that can handle up to 5 million transactions per month, Alva notes.
Customer service agents now can support clients who most need them. “Our agents used to handle 8,000 calls a day, but now that volume has dropped to 1,000 calls, and cross-selling has increased,” says Aisha Rahmi, E-Care Team Supervisor at Telkomsel.
He explains that agents can now deal with complex issues without the distraction of routine inquiries, which contributed to both employee burnout and long queues of waiting callers.
Worker Productivity: Global communications firms are embracing GenAI to dramatically enhance employee productivity, sales and customer service.
US-based Lumen Technologies, on a mission to unleash the world's digital potential by connecting people, data and applications, is redefining customer success and sales processes using Microsoft 365 Copilot. Lumen’s salespeople were spending up to four hours per week researching information for customer proposals. Microsoft 365 Copilot has cut that time to 15 minutes, saving these employees three hours of work per week. “The value of the three hours per week for their B2B sellers amounts to about $50 million a year,” Candiani noted.
By automating routine tasks, significantly reducing the workload on employees, and allowing them to focus on higher-value activities, Microsoft 365 Copilot has helped Lumen successfully transform its focus from traditional telecom services to being a technology-driven, innovative and efficient enterprise.
In addition, UK-based global telecom company Vodafone recently increased its Microsoft 365 Copilot licenses to 68,000 companywide, with the tool helping staff from legal teams to contact center operators do their work more efficiently. In the case of the contact center, Microsoft 365 Copilot helps employees summarize calls so they can better understand all the interactions the customer has already had with the company. The improved engagement has led to Net Promoter Scores – a measure of how much a customer would recommend the company to someone else – going from “low single digits into the high 30s,” said Scott Petty, Vodafone’s Chief Technology Officer.
Network Applications: Beyond voice platforms, GenAI also is creating a new breed of offerings that leverage the network like never before. For example, Azure Programmable Connectivity (APC) is a platform that provides a standard interface for developers to build apps that communicate with multiple mobile operator networks worldwide. It hides the specific complexities of various operators’ network APIs, ensuring cross-provider interoperability and reducing coding complexity for developers.
Another example is Yobi, a New York-based telecom and media company, which hopes to democratize access to scalable and secure GenAI-based data so telecom players can better understand customer behavior without compromising privacy. Yobi gives telecom and media clients access to advanced analytics and GenAI-derived insights into consumer behavior, network performance and operational efficiency, all based on the company’s own private-by-design data set.
AI-Driven Analytics: Finally, GenAI allows telecom firms to modernize operations by enabling new levels of real-time insights, intelligence and automation.
Yobi CEO Max Snow attributes GenAI for helping the telecom industry “reduce churn, increase revenue and improve customer experience—all without the time, expense or burdens of new data acquisition.”
Telecom operators can use GenAI to analyze customer behavior and network usage patterns, enabling more personalized service offerings, optimized resource allocation and improved marketing strategies.
Case in point: One of the key use cases of GenAI is the ability to optimize the daily routes AT&T’s field technicians take in their trucks to serve more customers and manage more repairs with less fuel consumption. The Ask AT&T platform, built with Microsoft, supports a handful of interesting early use cases, including improving the efficiency of care representatives who support customers, optimizing AT&T’s network and upgrading legacy software code.
AI has matured well beyond the chatbot craze and is now revolutionizing operations, providing better customer interactions and creating new product offerings. More than ever, the next generation of AI—GenAI—is how telecommunications operators will drive impactful results that maximize operational efficiency and network security and innovative new revenue streams. To learn more about how Microsoft is enabling telecom sector transformation through GenAI, visit Microsoft’s Telecommunications industry page.