• Nokia and STL have produced a report, Prioritising automation: creating a successful building block strategy
• Like all  so-called transformations the jump to AI-driven automation demands a culture change and upskilling

Intelligent Automation is the way ahead for Communications Service Providers. Having spent decades obsessing over how much more speed and capacity could be teased out of ‘the network’ by applying new technologies, the rising industry view is that, right now, there’s more value, more bang for the buck, to be realised  by getting the technology to do things for itself. Then, while it’s busy making data streams faster and more deterministic, network management more automated, energy consumption lower, human intervention less necessary, and opex much lower, the intelligence engines will also be able to gather and process network data to look after important aspects of the ‘business: things like customer experience management; fraud management ….. and perhaps most importantly, to help identify new Digital Service Provider opportunities. 

So the question is - if all this pans out, how much will ‘Intelligent Automation’ actually be worth to a CSP? ‘How long is a  piece of fibre’ answers won’t suffice.

Nokia, together with London-based STL Partners say they’ve done the research and run the data,  producing an extensive report which reveals that CSPs with an average revenue of $15 billion could generate an equivalent of 5.7% in annual value by applying  intelligent automation to facets of their business and operations, such as network and service operations, customer care, marketing and sales, and fraud protection.

Meanwhile Nokia and STL opinion  surveys reveal that  CSP executives are on-board with the  financial importance of incorporating intelligent automation into network management infrastructure.

Of the CSP execs surveyed, over 40% feel they have a clear automation strategy and over 70% noted that they currently also have a strategy to grow revenues from services that extend beyond connectivity and that they see automation as integral to delivering these services. 

“Automation is no longer a nice to have – it is essential. By incorporating intelligent automation, operators can not only better manage operational complexities, but can unlock revenue streams from new use cases across consumer and enterprise,” says Hamdy Farid, Cloud and Network Services Head of Applications at Nokia,

However Nokia’s survey reveals that there’s a long way to go. Despite the apparent enthusiasm, Fewer than 10% of the operators surveyed say they’re fully automated in any single network or service domain - the majority say they’re still in the mid-stages of automation. Importantly very few have achieved automation that spans multiple domains and therefore achieves the  inter-process operation that’s going to be key to maximising CSP ROI on Intelligent automation efforts.

How to proceed

Nokia and STL use the report to outline what they describe  as a building block approach to automation. STL points out that while rules-based automation has been around for decades and is well embedded, AI is at a more nascent stage with implementations in small pockets or proofs of concept (POC). So “automation” is an umbrella term encapsulating rules-based automation and AI and ML assisted automation.

So CSPs should adopt an incremental, use case-focused approach: defining and segmenting a key strategic focus area into practical and manageable blocks rather than embarking on a lengthy, all encompassing digital transformation programme. 

Operators should  instead identify opportunities for automation in standalone use cases and then group these into domains to form the building blocks. This  use-case by use-case basis allows  them to explore and understand how the technologies interact, but, warns STL, beware of constructing silos - when defining domains, operators must carefully consider how one domain will interface with another, as domain-to-domain interfaces are typically the most challenging part of the job. 

While operators have typically taken a monolithic approach to building applications in the past, there are clear benefits to building applications with microservices instead, says the report.  By leveraging cloud-native microservices and containers, operators can develop use cases as smaller, autonomous services, which can then be aggregated and scaled into working applications with greater speed and agility. A loosely coupled system gives operators greater control and reduces costs.

Above: Percentage to maturity: some use cases are getting there, but none are close to being fully automated 

People and Processes

That’s the framework, but the crucial ingredient is getting the right people into place and fostering the right culture (as is so often the case with any ‘next gen’ journey).  

There are ways to ease this process and its pain-points though. The reports suggests  integrating software developers and data scientists into business units,  but the report warns that this change process can be slow going for telecom operators. The survey shows that these changes in the people, process and culture for operators are slow-going and operators appear to be lagging on laying the groundwork - generally failing to adequately upskill their workforces to handle the new technical demands resulting from automation.

Only 20% of survey respondents said that their organisation has AI and automation training programmes in place for a wide range of employees, a development which is crucial. Network engineers who were previously engineering and operations focused will now transition to service management and acquire software development capabilities: instead of performing manual troubleshooting and reactive maintenance functions, network engineers will need to utilise their knowledge of the network infrastructure to build automated end-to-end processes. So a skills transition It is equally imperative for telcos aspiring to “grow connectivity services” or to “move beyond” to have clear personnel strategy and governance processes in place to support this skills transition; those belonging to the former segment tend to lag behind in this aspect. In addition, operators will need to balance who has responsibility for automation decision makingThe recommended approach is to employ a centralised team to focus on building skills and best practice in the more nascent automation areas (e.g., intent-based automation or machine learning) while at the same time allowing individual domains to identify use cases that would best in the medium term.

Download the Report here