AI lighting the path to the Dark NOC

Interview

BMC Helix’s Hector Villena discusses telcos’ autonomous network ambitions and how AI agents are unlocking hands-off Business Support Services (BSS)

As customer expectations rise and switching communication service providers (CSPs) becomes easier than ever, telcos are under intense pressure to differentiate through exceptional service and smarter operations. That pressure is driving a shift toward AI-powered automation, resulting in smoother service for customer and greater efficiency within CSP operations.

This shift is at the heart of BMC Helix ServiceOps, the company’s cloud-native platform for service management, operations, and automation. Launched back in 2020, the platform quickly proved popular, resulting in it being spun off as a separate company earlier this year.

“The landscape is changing very quickly, and flexibility is everything,” explained Hector Villena, Area Vice President for Sales in BMC Helix CSP Go-To-Market. “That’s why we decided to rearchitect the whole platform from the ground up, with a SaaS first mindset, open architecture, AI driven and completely interconnected. This means our customers can deploy on-prem or SaaS with the same capabilities.”

BMC Helix’s goal was to combine the company’s two strengths: Service Management and  Operations Management. Traditionally these two fields have been largely separated in the telco world, with data largely siloed off and unavailable for simple cross-analysis. With the AI powered tools from Helix, however, data from both worlds can be combined, resulting in a higher level of automation and efficiency.

“The biggest CSPs that we deal with have hundreds of thousands of assets, from servers and storage to network devices. Managing all of that is highly complex. We help them to visualise those assets and quickly assess how the components are related to each other. We then add our AI-driven BSS capabilities, analysing data, monitoring events, and providing a better understanding of that environment,” said Villena. “When you combine the ticketing information from the network with the data coming from events, metrics, logs, and telemetry, and use the topology of the services as a common map to link them, you have something really powerful”

Within the network itself, the benefits of these advanced AI tools are already being felt by customers. Using AI and advanced analytics to assess network data, networks can be made to better anticipate and prevent service disruption, improving uptime and avoiding network outages. When an incident does occur, the platform can also use AI to identify and diagnose the issues behind it and even automate remediation.

“You have millions of tickets going through those platforms. We collect data about previous incidents and their solutions, then apply AI to help deliver a resolution,” said Villena. “For starters the platform can quickly correlate and identify those events and incidents that are essentially related, reducing the number of tickets dramatically. Some of those resolutions can be automated, but for those that can’t, it’s still providing a major boost in efficiency. It’s providing the data and instructions so a Level 1 technician can handle the issue much faster. The initial results we are seeing in our customers are outstanding: 70%-86% Qualified event noise reduction within days, MTTD reduction by 90% or MTTR optimisation over 55%. I understand such a range of improvement is hard to believe, but I do invite any CSP struggling with these challenges to give us a call and allow us to understand their process and prove how the platform could deliver these transcendent improvements to them.”

A growing role for AI agents

AI agents – AI algorithms specifically designed to automate workflows through problem solving and decision-making without human intervention – are playing a major role in this digital transformation. CSPs are already beginning to deploy these agents to automate various customer journeys, helping provide support to staff in call centres or create bespoke packages for B2B customers.

But, for Villena, AI agents’ biggest strength could be their use for internal CSP operations. To date, BMC Helix has released 12 agentic AI agents, focussing on areas of telco operations that can provide the biggest gains in efficiency. The ‘Employee Navigator’ agent, for example, serves as the first layer of AI interacting with the user in the Helix platform and can handle simply administration tasks for employees.

“If an employee needs to book annual leave, for example, they can talk to the Employee Navigator prompt in plain language, and the prompt will ask for the data it needs (if any) to clarify the situation, and then it will book the leave automatically. The employee doesn’t need to manually go through a complex HR system,” Villena explained. “We’re trying to stop employee questions from ever reaching a human agent if they don’t have to, so everyone has more time to focus on higher value tasks.”

The agents can also impact event resolution itself.

“A technician can, thanks to HelixGPT, ask an agent for a problem classification synopsis, a brief root cause summary with the contextual information (metrics, events, behaviour of the system, etc.) and actionable insights. Following that, the technician can ask the system to provide step-by-step recommendations guide to solve the issue. Or, if the issue is so complex it requires the involvement of different subject matter experts, it will automatically create a Teams group chat pulling in the right resolution engineers where all relevant stakeholders can ask questions to the agent (to get a full 360º on the situation, or root cause analysis) and can work with each other to solve the ticket. These are just few of the use cases that can be delivered by the BMC Helix platform” said Villena. “It’s optimising not only the Mean Time to Repair, as it brings the resolution to L1, but also reducing the overall cost per ticket”

BMC Helix is planning to expand their roster of AI agents in the coming months to cover even more use cases, helping to empower telco workforces even further.

Is the ‘Dark NOC’ within reach?  

With more and more AI agents within the telco network, just how far can telcos take automation within the Network Operations Centre (NOC)?

The NOC serves as the nerve centre of the network, the physical location from which network activity is monitored and managed. Currently, most NOCs are highly manual, with network engineers overseeing data traffic and the status of the networks vast physical and digital assets.

As AI and automation becomes more sophisticated, however, the need for manual, human intervention in the network is decreasing. This naturally leads us to the concept of a ‘dark NOC’, a fully autonomous NOC that leverages AI and machine learning to manage the network without the need for any human oversight at all. This, says Villena, is the end-goal for most CSPs.

“CSPs are trying to work smarter and as a result automate as many operations as they can. A consistent objective for many of them is to reach ‘zero touch’ operations, where the operator does not interact manually with the network at all. Network incidents come into this big AI brain and then get identified, triaged, root caused, and resolved automatically. That’s the aspiration of all CSPs.”

But just how far are we on this journey towards fully automated NOCs?

“We’re still a long way from [the dark NOC]. But each part of the network will have a different level of automation, with some already being highly automated,” said Villena. “There’s a big difference between the level of autonomy telcos have in their network functions and their wider IT architecture.”

“In terms of their IT architecture, most telcos I talk to categorise themselves as between Level 0 and Level 1,” he added, referencing TM Forum’s framework of Autonomous Networks Levels, with Level 0 being fully manual and Level 5 being fully autonomous. “Our aspiration is to enable these telcos to get to Level 4 within a short period of time”

But rapid advances with AI and related technology means progress towards greater automation has been fast.

“Some CSPs are aiming to have 80% of network incidents automated by the end of 2026, and to achieve full autonomy by 2030. That’s a very aggressive timeline, but the aspiration is clearly there,” said Villena. “AI has really enabled that transformation. We’re seeing some incredible results in our platform which really show this could be achieved in the next years. The level of improvement is enormous.”

Part of this acceleration relates to the increasing convergence of IT and telco functions.

“As networks are becoming more virtualised and cloudified, the division between telco and IT systems is becoming blurred,” said Villena. “Part of our ‘Operations in the future’ strategy is to enable and integrate with those systems that exist on telco and in IT. With this, we’ll be able to help our customers move from Level 0 and Level 1 to Level 4 and Level 5 autonomy very quickly.”

The building blocks for even greater autonomy

While for now achieving the dream of a Dark NOC remains tantalisingly out of reach, there is little denying that the convergence of CSP systems and the introduction of more sophisticated AI is making rapid progress towards this goal. AI-driven platforms like BMC Helix are proving critical in bridging the gap between legacy complexity and autonomous efficiency, helping to deliver better experiences for consumers and major cost savings for operators.

Also in the news:
US judge rules Huawei must face charges of fraud and racketeering
Optus ditches football rights to focus on telecoms
Nokia launches digital twin platform Enscryb to digitalise energy sector