When was the last time you called your telecoms provider and actually walked away feeling satisfied, asks Anna Ribeiro? If you’re struggling to remember, you’re not alone. For years, call centres have been a pain point, plagued by long waits, robotic menus and agents reading from tired scripts.
But lately, something’s shifting. Have you noticed that responses feel quicker, smarter, even a little more human? That’s AI at work behind the scenes. And telcos are leaning into it by using AI to rethink how they deliver support, cut costs and retain customers.
AI adoption in telco call centres
The global IT and telecommunications call centre AI market was worth about US$302.9 million in 2024 and is projected to surge to US$1.005 billion by 2030, growing at a CAGR of roughly 22%. This jump mirrors a broader AI backswing across industries, as telcos also see automation as a critical way forward.
Market research firm Technavio forecasts that the global call centre AI market will likely reach close to $4.3-billion with a CAGR of around 27.2% from 2024 to 2028. Data from the IMARC Group estimates that the call‑centre AI market will grow from US$2.3 billion in 2024 to roughly about US$12.8 billion in 2033, recording a CAGR of about 21.3%.
While a few telcos were early to modernise their data ecosystems by building capabilities, products and digital twins for core areas like networks and call centres, McKinsey points out that most are only now beginning to use generative AI as a catalyst to upgrade their data infrastructure. This shift is pushing them toward hybrid lakehouse architectures and structured data products designed for reusability and consistency across multiple use cases.
Analyst insights and market trends
As AI assumes routine interactions, 65% of call centre interactions are now automated using AI, Gitnux’s April report disclosed, while call centre volumes have declined by as much as 30% as a result of these technologies. In addition, AI-powered personalisation has driven 15% more customer satisfaction, 35% fewer false positives in fraud detection and 20% more efficient networks. Meanwhile, virtual agents now cover more than 60% of technical support requests, accelerating service while taking the strain off human agents.
The trend around AI at telco call centres is not so much about an acceleration but more of a strategic shift. Analysts now predict that three in four of the new contact centres set up by 2028 will include generative AI, mirroring the technology’s increasing maturity and importance. With this image of a fast, smarter and more human touch support becoming the new norm and vendors are increasingly focused on integrating AI techniques and humans to bring about a transformed future of customer experience. Through careful design, trust and oversight remain necessary, telcos that apply AI are clearly laying the foundation for long-term resilience, efficiency and greater customer adoption.
An arXiv study found that AI assistance causes an average 15% increase in worker productivity in the number of issues resolved per hour, with significant heterogeneity across workers. Meanwhile, the less skilled are getting faster and better and the lowest skilled are catching up, while the most skilled are getting a little faster and a little worse. Moreover, AI induces greater customer politeness and fewer requests to ‘speak to a manager.’
What vendors bring to the table
A 2024 Verint study of 300 contact centre leaders in the US, UK, Australia and New Zealand revealed that 53% see AI as the most important driver of customer experience automation today, compared to just 10% who think traditional telephony is the strongest force in the UK. Approximately 80% of respondents believe that the ability to incorporate AI, automation and bots for self-service and agent-assisted interactions is important, yet only 41% are satisfied with their current capability and 46% are satisfied with this capability within their company.
While telcos are keen to move forward with AI and machine learning in customer engagement, there is work to be done and vendors can step in and fill the gaps in capability and integrate it into the broader OSS/BSS environment.
NICE crossed US$1 billion in annual cloud revenue in 2021 and by March 2023, it had exceeded one million users on its CXone platform. The company established itself as the forefront of the call centre revolution in 2023 when it unveiled ChatGPT-powered conversational AI for CXone. The 2024 release of CXone Mpower united the integrated suite of solutions, such as Copilot, Autopilot and Actions, for end-to-end automation throughout the customer journey, including front- and back-office processes. This year, the firm unveiled its Orchestrator update, which combines agentic AI with human collaboration to deliver end-to-end AI automation.
Uniphore powers worldwide conversational automation, enabling up to 75,000 customer-service agents to handle nearly 160 million interactions a month. The fact that it appears in telecoms contact centres is a fine example of operational usage and adaptability across important points of contact with contact centre operations. The company markets its technology as a way to boost and support the efforts of call centre staff, not replace them.
As each of the vendors adopts different techniques of embedding AI into telecoms customer service, they highlight how quickly AI is shifting from an experimental tool to more of an operational cornerstone for telcos looking to modernise and future-proof the way they do things in the call centre.
Couple of caveats to keep in mind
Telco call centre AI offers efficiency and personalised service but brings challenges. Data bias and lack of explainability can erode trust, while privacy and security risks call for greater compliance. Integrating into legacy is expensive and there is a shortage of talent to deploy. Over-automation will alienate customers and cultural tools such as accent neutralisation are already sparking ethical debates. AI workloads at scale have added environmental cost.
Telco use cases cementing AI’s role
Verizon uses generative AI to predict call reasons 80% of the time, matching callers with the best agents and cutting store visit time by around seven minutes, in a bid to retain an estimated 100,000 customers this year. AT&T, handling 40 million calls annually, moved from ChatGPT to an open‑source AI suite fine‑tuned for its needs. The hybrid system retained 91% of prior accuracy, slashed AI costs to 35% and cut processing times from 15 hours to under five hours daily.
Global digital business services company Teleperformance is trialling AI from Sanas that ‘neutralises’ Indian and Filipino accents in real-time to improve clarity and reduce handling time, while preserving the speaker’s emotion and identity. This will work alongside background noise cancellation technology to improve call quality.
In conclusion, the adoption of AI technologies across telco call centres is no longer an abstract experiment; it has become a part of the day-to-day operation. Improvements in speed, accuracy and personalisation are just too valuable to row back on, while the competitive imperative to offer frictionless service does not offer much space to step back.
While there are ethical, technical and cultural barriers that remain to be crossed, the direction is obvious. AI will continue to deepen its role as it reshapes the customer experience and the job of the human agent. Now, for telcos, the question is not if they should turn to AI, but how deep and how fast they need to go.
Anna Ribeiro