Skip to main content

How to solve the automation puzzle

Practical insights for CSPs

Ask Telenor Senior Research Scientist Pål Grønsund why automation matters for communications service providers (CSPs) and he won’t mince words: the 5G vision is not possible without it.

“The 5G vision is to serve multiple vertical industries that have many different requirements,” Grønsund says. “That means slicing, and when you implement multiple slices, you need automation. Whatever you want to achieve with 5G, I would put automation all the way at the top in importance.”

Grønsund would know. He is the coordinator of the European Commission’s 5G-VINNI (Verticals Innovation Infrastructure) Consortium and Telenor is one of 23 partners experimenting with multinational 5G networks. Automation is one of the key objectives of the project — reflecting the industry view that networks won’t be able to deliver next-generation services and risk being overwhelmed by increasing complexity without it.

And yet automation is a complex challenge itself: a giant puzzle with thousands of pieces. As CSPs pour time and resources into automating their networks and businesses, many are dogged by uncertainty: Are we doing it right? In the right places? Could we be doing it better?

Providers under pressure

In February 2021, Nokia and STL Partners surveyed 100 CSP executives and found cost control, customer experience enhancement and service innovation were the top drivers of automation efforts. External pressure from hyper-scalers, data center providers and device and application makers are other key factors, according to Chris Lewis of Lewis Insight.

“These external parties are demanding the network be simplified. They’re demanding fixed and mobile broadband along with responsive services, and they want those services deployed in a much shorter timeframe than in the past,” Lewis says. “Automation will change the way services are created, delivered, managed... pretty much the whole business. It’s fundamentally different from the way we've run telcos in the past.”

companies aspiration

CSPs’ own shifting ambitions are also making automation increasingly critical. In the Nokia–STL survey, 71 percent of respondents said they aspire to move “beyond connectivity” aiming to offer services that build on value-add platforms and applications. That inherently requires business, network and service automation.

Most CSPs still operate with a mix of automated and manual processes. About 40 percent of survey respondents said they have some degree of “intra-process” automation (automation within a single domain), while few have achieved “inter-process” automation across network and service domains or business units. This more sophisticated kind of automation is exactly what’s needed to bring disparate systems together in a seamless end-to-end whole.

Four principles of automation success

Nokia reached out to a diverse panel of industry experts for their take on what CSPs could do to make solving the automation puzzle easier. While each brought a unique perspective, the experts agreed four principles are key: having the right mindset, knowing why you’re doing it, being patient as you work incrementally, and embracing openness.

Principle 1: Change your mindset

“Everybody immediately thinks automation is a technology problem,” says George Glass, Chief Technology Officer at TM Forum. “But it's actually a deep cultural issue. It's a people issue. It's a skills issue. You're basically moving people into other people's territory, bringing disruptive thinking, asking people to disrupt themselves. A transformation like this is 90 percent people and culture and 10 percent technology.”

Chris Lewis agrees. “The culture change cannot be overstated. It's a massive change to get the IT department to work more closely with the network department to work with the lines of business, to the product, people and so on.”

Our automation experts

George Glass

George Glass
CTO,
TM Forum

Pal Gronsund

Pål Grønsund
Senior Research Scientist, Telenor

Chris Lewis

Chris Lewis
Industry Analyst, Lewis Insight

Tilly Gilbert

Tilly Gilbert
Senior Consultant, STL Partners

Ajay Singh

Ajay Singh
Head of Advanced Services, Nokia

TL Viswanathan

TL Viswanathan
Head of Digital Operations, Nokia

Because automation demands a culture shift, CSPs need a plan to go about it — not just a technology roadmap, but also a clear people roadmap for automation and AI.

“It’s a people issue. It’s a skills issue. A transformation like this is 90 percent people and culture and 10 percent technology.”

George Glass, TM Forum

The kind of out-of-the-box, innovative thinking needed to execute a comprehensive automation strategy will emerge only if a CSP has the right organizational culture to generate it. Shifting the culture is necessary but by no means simple. It demands forethought, dedicated effort and careful management.

“Leadership is an extremely important part of making this transformation,” says TL Viswanathan, Head of Digital Operations at Nokia. “It all starts there. After all, it’s the leaders who decide, ‘Why are we doing this?’”

Principle 2: Know your ‘why’

One of the most important items to get clear on is the business purpose of automation, Viswanathan says. “It's not about what you can automate. It’s about why you should automate it. It’s about the benefits it will bring.” To yield the most meaningful benefits, whichever use cases are chosen, they should be value-driven.

Yet, as the Nokia–STL survey results show, the ‘whys’ of automation are still largely tactical, with some variation by region: cut costs, speed up time to market, manage complexity and launch new services that can’t run without automation. That’s something CSPs should look to change, Chris Lewis says.

“Yes, we have to automate the network, but we also have to automate all the pieces that fit around the network to deliver the service that, ultimately, customers will pay for. I think we've been too obsessed with the underlying technology rather than the service that’s being delivered or the economic value being created.”

The ability to focus on value creation requires a clear automation strategy and a formalized process to keep goals in view. Neither is especially common, despite CSPs’ own impressions of the approaches they’re taking. One finding from the Nokia–STL research is that while 40 percent of CSPs say they have an automation strategy, only one in five systematically tracks key performance indicators (KPIs).

driving force

Closing that apparent gap between strategic intent and implementation would not only help CSPs measure their progress but also structure their automation ambitions in achievable chunks and avoid taking on too much at once. The experts on the panel generally agreed that it’s a good idea to start small and focused and scale from there.

 

“I think we've been too obsessed with the underlying technology rather than the service that’s being delivered or the economic value being created.”

Chris Lewis, Lewis Insight

Principle 3: Work incrementally

“We should not try to solve automation as a 1,000-piece puzzle,” Viswanathan says. “We should first try to solve it as a 10-piece puzzle, then a 20-piece puzzle and then a 30-piece puzzle, and start to build organizational maturity to handle automation, and scale it. If you start with a 1,000-piece puzzle, you’re bound for failure compared to starting with a smaller problem set.”

Proceeding incrementally ensures real progress is made and revenues don’t ‘leak out the door’ because of inaction. Once CSPs have deployed basic automation, they can move on to deploy more complex use cases such as network slicing and “as a service” models. This is similar to Abraham Maslow’s famous pyramid, which showed that human beings must meet their basic physiological needs before addressing more complex ones such as self-esteem and self-actualization. Automation builds domain by domain and layer by layer, delivering richer capabilities and benefits with each progression.

Automation domains - a hierarchy of needs

automation puzzle

In the automation hierarchy of needs, the bottom-layer functions focus on efficiency requirements — dealing with the abundance of data and processes that come with deploying a massive network such as faults and alarms, Viswanathan says. “Those are all a construct of network and software automation, which has become a fairly basic requirement.”

The top of the pyramid, in contrast, is all about business requirements. “For example, crashing the time to value of services is a huge requirement today in the 5G era,” Viswanathan adds. “Automation in the service and customer management domains needs to drive that outcome.”

Climbing up the pyramid invites increasingly sophisticated automation use cases enhanced with artificial intelligence — capabilities that can’t be coded into software and require a new cycle of automation, according to Ajay Singh, Head of Advanced Services at Nokia.

“Hyper automation will enable tasks that are completely hands-free or closed loop, that don’t need human operators,” Singh says. “Given that AI science and technology are just developing, especially in the telco world, this second ‘S curve’ of hyper automation is just starting now.”

CSPs should be as pragmatic and incremental about climbing the pyramid as they are for their other automation pursuits, Glass says. Break the challenge down into smaller parts so you can succeed: “When you stand back, you have a much richer picture to look at in the end.”

Once enough discrete processes have been automated, it’s then possible to start connecting the pieces together — provided they’ve been automated in a consistent, standardized and open way.

Principle 4: Use open APIs everywhere

Embracing openness — specifically, open interfaces — is key to linking up the pieces of the automation puzzle.

“Open APIs make it easier to connect bits of data, functions and business processes together in a frictionless manner,” Glass says. “If you want to automate a set of components, you have to standardize the components first.”

That standardization of a common set of interfaces has to go all the way, because half measures won’t deliver the full benefits possible. The industry saw something similar with network function virtualization (NFV): vendors virtualized functionality and put it on standard infrastructure, which delivered some cost savings, but most networks still had vendor-specific operation management and control interfaces, Glass recalls. An industry-standard set of open APIs for automation would make it possible to automate more across those components.

“If you want to automate a set of components, you have to standardize the components first.”

George Glass, TM Forum

“Standardized APIs are very important,” Grønsund agrees. “And having a multi-vendor environment, it actually forces you to have it. If not, it will be difficult.”

With so many domains to cover, no single industry body could be expected to develop a complete set of standardized, open APIs from end to end. In Glass’ view, having multiple parties contribute is the only path to achieving a comprehensive, standardized approach — provided those parties work collaboratively toward a common goal.

An invitation to ‘be bold’

Taken together, the four principles of focusing on mindset, knowing the ‘why’, working incrementally and using open APIs equip CSPs with a flexible approach for pursuing their automation goals. CSPs should use all the imagination they can muster to determine what those goals ultimately ought to be, according to Glass.

“It goes back to Henry Ford's analogy when he was inventing the motor car,” he says. “If you asked people in those days what they needed, they would have said faster horses. And CSPs have done something similar: they've pursued a strategy of trying to get faster horses by automating their existing processes instead of taking a step back and saying, ‘Can we actually transform our business and remove the need for some processes altogether?’”

This ties directly to the business question of whether CSPs want to grow purely as connectivity providers or grow beyond connectivity and seize the full possibilities of the 5G era. The latter requires the motor car approach, with rigorous automation interfaces between traditionally siloed business units. Faster horses won’t cut it.

“You can’t supplant the 4G operating model and hope to run even a basic 5G operating model that captures the promise of the enterprise market and differentiated service-level agreements,” says Singh. “You need what we would call ‘hyper automation’ — completely hands-free, closed-loop automation. Otherwise the 5G business case will fall flat.”

The experts on the panel all advocated for making the bolder choice because committing to a simple evolution of the status quo today makes it harder to justify bigger shifts down the road, and opportunities could be missed. Thinking of automation as an instrument of business and cultural transformation will unlock the full possibilities of 5G.