Artificial Intelligence in Telecommunications 2025

These agents surpass static, decision-tree chatbots by offering human-like, context-aware interactions that continuously learn and adapt based on real- time customer behaviour and organizational data. They provide seamless, personalized interactions by adapting to changing customer needs, using signals and integrating both structured and unstructured real-time data from all customer touch points. As a result, CSPs are significantly increasing adoption rates to the extent that they are now among the top three channels preferred by customers.19 Technologically, this marks a shift from traditional, fixed ML models trained on historical data to sophisticated LLMs pre-trained on customer interactions. These models are further enriched with broader market and organizational context, enhancing the intelligence and adaptability of AI- powered customer engagement. GenAI also improves the efficiency and service standards of human-assisted channels by harnessing and democratizing knowledge. This empowers workers to become “super agents” capable of cross-functional support. This reduces call handling times, lowers service costs and unlocks opportunities for incremental sales. As intelligent automation advances and data foundations evolve to integrate customer data across sales, marketing and service, the traditional contact centre – seen as a cost centre handling transactional interactions – can be reimagined. It can become a unified, proactive and predictive service hub that integrates all channels, enhancing customer loyalty, driving in-life sales and supporting business growth.Using its natural-language capabilities, genAI can enable always-on, real-time digital companions that provide 24/7 service, referring complex issues to human agents when required. Secure and reliable operations The evolving security landscape requires telcos to adapt to the new opportunities and threats that AI brings. AI’s reliance on extensive datasets in large- scale compute and storage environments creates a complex and attractive attack surface, making defence increasingly challenging. Open interface networks and traditional IT technologies such as cloud, hypertext transfer protocol (HTTP) and application programming interfaces (APIs) expose network infrastructure to a broader array of threats,20 while legacy components, often retained due to capital constraints, amplify vulnerabilities. Telcos handle sensitive data from users, such as location information and communication content, which are prime targets for state-sponsored attackers and cybercriminals. Moreover, genAI is being weaponized to deploy malicious code at scale, enabling faster and more extensive damage. Managing these risks requires advanced AI capabilities. AI can identify and patch vulnerabilities and analyse vast operational datasets in real time to detect security incidents and prevent fraud. By automating heavy-lift tasks, AI enables security teams to focus on tasks that engage their human security skills, creativity and teamwork. Key AI techniques for securing operations include network planning, adversarial testing, model evaluation, pattern matching and user behaviour analysis. AI can identify and patch vulnerabilities and analyse vast operational datasets in real time to detect security incidents and prevent fraud. Artificial Intelligence in Telecommunications 9