There is a lot of hype surrounding Ai and how it is set to revolutionise the communications and service provider sector.
An incremental approach to adoption is most likely. But where are we now? Artificial Intelligence (AI) and Machine Learning (ML) represent the next technological evolution for communications. Even further, many CSPs view it as a communications revolution that will provide significant and multiple benefits, enabling them to manage and optimise increasingly complex networks, provide dynamic orchestration of services and network slicing, offer zero-touch network fault recognition and resolution, and so much more.
The benefits of AI and ML to CSPs
Here are just some of the perceived potential benefits of AI and ML:
- Improved customer experience
- Optimised network operation and maintenance
- Improved decision making
- Lower costs and operating expense
- Increased sales
- Reduced customer churn
- New service development
No wonder there is so much hype about AI in the telecoms sector. But how far down the path are we in terms of lighthouse projects and deployments? And when can we expect to see the delivery of the myriad benefits it promises?
A survey conducted by the TM Forum in June provides some significant insight into where the industry stands when it comes to the deployment of AI solutions[1]. The survey questioned 141 individuals from CSPs around the world and builds on its autonomous networks (AN) Benchmark report, published in December 2024[2].
It describes six levels of AN (from 0 to 5), with levels 4 and 5 inherently requiring AI and ML to achieve their goals.
How far down the line is AI deployment in telcos?
The same report suggests that only 4% of respondents had implemented Level 4 AN (including AI) in any form – and this is likely to be on a task-based basis.
No CSP or operator has yet embarked on an enterprise-wide AI deployment, with many citing lack of budget, a lack of clear business objectives, cross-domain integration challenges, and no clear path to reach Level 4.
However, 23% of respondents to the TM Forum report believed they would reach some Level 4 deployments by 2026, with that figure growing to 85% by 2030. Clearly, eyes are on the longer-term objectives. The report notes: “A handful of CSPs have publicly stated they are adopting AN Level 4 ‘high-value scenarios’, aiming to reach Level 4 for operational flows in some domains from 2025 to 2027.”
However, this still points to a piecemeal approach, with automation and AI building over the next 5 years, with ‘low hanging fruit’ applications, and single domains, likely to be targeted first. In turn, these can be integrated over time, reducing risk, cost, and disruption, to create cross-domain capabilities.
It means that different functions and domains across the network are likely to evolve at different paces, with some more advanced from an AI perspective than others. This is a scenario borne out by GSMA Intelligence, which has found in its own research that there is a gap between deployments of core AI — AI used to drive operational processes — and Generative AI (GenAI), which delivers new content.
Core AI will be integral to new processes, while Gen AI unlocks new modes of interaction. As a result, GenAI brings significant benefits for enabling customer care – so it’s no surprise that the GSMA has found this to be much more widely deployed than Core AI, as part of its ongoing series on AI.
In the most recent edition of the GSMA’s research (Telco AI: State of the Market, Q3 2025), Customer Care together with Sales & Marketing accounted for more than 60% of current AI deployments.
On the other hand, because of the potential Core AI offers for enhancing operational efficiency and performance, the situation is reversed, with many more respondents reporting that they have deployed Core AI across “a few” business areas. And, in any event, Core AI has long been the target for many – but GenAI has swiftly taken the lead.
In other words, delivering Core AI is likely more difficult and will take time – so GenAI represents lower hanging fruit for rapid returns. And, perhaps is seen as less risky than deployments in core operations.
How does We Are CORTEX help on your automation and AI journey?
It’s also borne out by our experience at We Are CORTEX. The majority of customers view customer experience, optimisation of network operations, and boosting employee efficiency and productivity as the important drivers.
But, most important, is to realise that AI will build on existing automations and what you already have.
We believe that AI should be added incrementally where it solves problems, enhances value, or delivers new benefits. We also maintain our focus on maximising return from existing and ongoing investments, and reducing risk. We focus on four key areas:
- LLMs in process flows
- Assisted processes
- Design Co-Pilot
- AI agent integration
We are helping CSPs right now to build automation and AI roadmaps that utilise all existing automations, so nothing goes to waste and to minimise the risk of deployment.
To find out more about how we can help, download our latest paper by clicking here.
[1] https://inform.tmforum.org/research-and-analysis/reports/a-regional-guide-to-autonomous-networks-progress
[2] https://inform.tmforum.org/research-and-analysis/reports/autonomous-networks-in-search-of-best-practice
![AdobeStock_494624009 [Converted]](https://www.wearecortex.com/wp-content/uploads/2025/12/AdobeStock_494624009-Converted-960x718.jpg "AdobeStock_494624009 [Converted]")
