AI, data, automation: the blueprint for profitable autonomous networks

Charles Bligh, CEO, Mycom

---

Communications service providers (CSPs) are operating under structural pressure with 50  to 70 per cent of their revenue consumed by opex. They have high amounts of trouble tickets, long resolution cycles and large engineering teams to manage growing network complexities. At the same time, CSPs are targeting Level 4/5 autonomous networks (as defined by TM Forum) within five years – ahead of the 6G wave.

 

The reality is that manual operations cannot scale and incremental efficiency gains are no longer enough.

For CSPs, constant improvements to network availability, quality and customer experience are fundamental to competitive advantage to drive growth. The sustainable way to keep improving is with automation. Automation in CSP networks prevents issues before they become faults, and when issues occur automation offers immediate response. A move towards higher levels of network autonomy can reduce opex significantly. This means that CSPs need high-quality, real-time data combined with AI and automation to deliver real business benefits.

 

But automation is not easy to achieve. Successful CSPs have eliminated silos of data, and sustained focus on data specifically for network quality. In addition, their operations and IT teams have joined forces to ensure data consistency. In essence, they have built solid foundations for data quality which makes automation much easier to drive in real time.

 

To further this effort, they have adopted AI technologies which can lower cost and create better customer experience, enabling growth. However, this is an ongoing journey for most CSPs. As per Gartner, 57 per cent of organisations estimate their data is not AI-ready. Organisations without AI-ready data will fail to deliver business objectives and open themselves up to unnecessary risks.

 

Dealing with the vast scale and complexity of modern telecom networks

 

As networks rapidly grow in size and complexity, the only way to manage them cost-effectively is with automation. Creating an autonomous network requires data to fuel it. However, inconsistent or incomplete data leads to unreliable decision. It’s important to note that automation amplifies data weaknesses rather than correcting them.

 

AI has opened up more possibilities in automation, but accurate data is critical for this. The CSPs who have spent time working on this in the past five years will power ahead, and those that haven’t must quickly pivot or risk being left behind by their competitors. Gartner says that to scale AI, leaders must evolve data management practices and capabilities to ensure AI-ready data – determined through the data’s ability to prove its fitness for use for specific AI use cases – can cater to existing and upcoming business demands.

 

A data analytics platform that fuels high-level AI and automation

 

To resolve some of these challenges, it is imperative that CSPs build data analytics platforms, which can, at vast scale, ingest network data in real time from a multi-vendor network, standardise the data, add context to it for consistent interpretation and produce high-fidelity outputs suitable for AI and automation.

 

Basically, the data analytics platform would remove all ambiguity and inconsistency that would otherwise undermine AI results and becomes the single source of truth in real time.

 

The automation maturity model

 

Mycom’s three-tier approach aligns AI capability with automation maturity and uses the following three technologies:

• Rule-based
• Machine learning
• Generative AI including agentic AI

Today, many CSP networks operate at Autonomous Networks Level 1-2 (as defined by TMForum) and some CSPs have reached 3.X, using rule-based automation, which is fast and cost-effective. Machine learning is proven and, at scale, can be very cost-effective and highly accurate. Finally, generative AI or LLM-based technology is relatively new and CSPs are investing and deploying its agentic AI flavours to deliver value but they are still high cost and accuracy needs to improve, but it will improve quickly.

 

As the technology matures over the next five years, cost will come down and accuracy improves, automation will be increasingly driven by ML and LLM. Furthermore, agentic AI can perceive environments and act to achieve specific outcomes by collaborating between multiple AI agents.

 

Mycom recommends a use of all the above technologies, depending on the stage of automation of the CSP.

 

The agentic AI imperative

 

The agentic AI market is very promising: Deloitte predicts that the Agentic AI market may reach $45 billion by 2030. With agentic AI adoption, CSPs can bring in corporate-level transformations as well. BCG suggests 20 per cent to 30 per cent faster (auto) resolution of IT service tickets and 60 per cent financial risk events reduction in pilot environments. Gartner says that by 2028, 60 per cent of brands will use agentic AI to facilitate streamlined one-to-one interactions. And that by 2029, agentic Al will autonomously resolve 80 per cent of common customer service issues without human intervention, leading to a 30 per cent reduction in operational costs.

 

This is not incremental improvement; it is structural margin expansion.

 

At Mycom, we are already seeing signs of this shift. As accuracy improves with the data getting better and cost reducing, use of agentic AI for automating key network engineering tasks has already begun. Mycom AI agents are now collaborating with other OSS AI agents to resolve problems faster, akin to a cross-functional team of experts to solve problems.

 

Leveraging agentic AI for enterprise intelligence

 

Enterprise is where the real value of agentic AI can be magnified. It can drive better decision-making use not just for the network but also for customer experience data, churn data, market data and support business decisions proactively.

 

The good news is that CSPs are uniquely positioned for agentic AI because of their existing mature infrastructure and data-rich environments.

 

Network data has traditionally been filtered through technical teams, which offers limited transparency for business leadership; almost a technical “black box”. Now, with AI in place, data can be exposed as insights to create value, such as bespoke offers for customers or improving customer service. Streaming data at scale in real time to data lakes and other systems enables this insight in real time. At Mycom, we have invested in generative AI to provide direct conversational access to network intelligence and customer data so that executives can query performance using natural language.

 

The future of AI and autonomous networks

 

Today, only about 30 per cent of CEOs are satisfied with the ROI of AI as they struggle to find AI use-cases and talent. It is time for AI-native software engineering to make its debut, leading to AI-native networks. To get there, the first step is for CSPs to fix their data foundations.

 

CSPs are targeting Network Autonomy Level 4/5 over the next five years, and we need to get this right before the next big 6G wave.

 

Trustworthy AI inputs will help CSPs achieve high levels of automation at the required scale. Mycom is evolving along these trends and has transformed itself over the past several years from a network/service assurance company into a cloud-enabled, AI-driven data management and automation partner. We are excited about this next wave in the industry and the growth it will drive.

 

The focus on AI, data and automation is not about deploying AI tools. This is about redesigning cost structure, unlocking growth through superior experience, achieving Level 4/5 autonomy and building defensible competitive advantage.

---

For more information, visit mycom.com

---

Charles Bligh, CEO, Mycom