Tupl Co-Founders on How Wireless Networks are Becoming Increasingly Automated

Automation has already been embedded in wireless networks for quite some time with operators trying to deploy solutions that would help them be more efficient.

Automation has already been embedded in wireless networks for quite some time with operators trying to deploy solutions that would help them be more efficient. There are examples that go all the way back to frequency planning or cell-planning; however, these solutions are quite outdated for the state of mobile networking today. That’s why we founded Tupl in 2014 — we set out to develop a solution that can address the growing need for automation tasks that handle complex manual work happening day in and day out in network operations.

In recent years, real technology breakthroughs have resulted in the techniques in machine-learning that are enabling this automation to go one step beyond what was previously possible. The speed to create automation solutions is faster, and we can go deeper than regular coding could before. We can capture things that are in more gray areas. And as we have seen, machine-learning and deep learning are taking over in many industries, not only telecom.

Motivations for Automation

Automation, AI, machine learning and deep learning are taking hold across all industries today. We see this happening in the wireless industry because the complexities of network operations are increasing dramatically. This is due to the growing size of the networks and the volume of components being deployed, resulting in more sales and more IoT enabled devices

At the same time, customers are becoming increasingly demanding of quality. Operators need their networks to become much more efficient, and the only way to achieve this at scale is through automation.

While we see the adoption of automation technologies increasing in wireless this year, we also see this as a long-term journey that is gradual. As operators embrace this transformation, they’ll begin to see benefits from day one, and then invest more effort into building out their vision for automation solutions.

Cultural Disruption and Adoption

Improvements in speed of decision making, customer experience and operational efficiencies are all a part of adopting automation technology that one would expect – there will also be a cultural shift in how companies and their people operate within their roles. Engineers and managers may fear for the future of their roles with automation coming into operations, but this has not proven to be the case when operators embrace the bleeding edge of automation by AI. Instead, there’s greater opportunity for managing these automated systems and providing more strategic value.

For example, when an operator deploys an automation solution for a particular use case, the system should adapt to the existing processes with the operators’ own engineers overseeing the models’ creations. Engineers will thus maintain oversight and control of these deployments, and once deployed, the engineers can begin adding new value and new functionality that wasn’t previously possible through manual processes.

Each operator is also different in how they will deploy automation solutions since every operator has their own prioritization and flavors of current pain points. For instance, fast growing operators may need to find efficient ways of augmenting their workforce by leveraging solutions that design and optimize the network faster or troubleshoot product issues more quickly. Operators focused on customer experience will want to see their customer care processes and ratings improve through automation. In several studies, it has been noted that only 1 out of over 20 customers actually reach out to complain, thus it is paramount to resolve issues that may result in unreported customer dissatisfaction and potential churn.

The cultural change occurs when the operators and engineers break down both technological and departmental siloes as well as old processes for the sake of creating new, more efficient ones. We see the future of telecom engineers shifting from manual day-to-day work into a higher-level path as they become software and systems designers. They will focus on handling all these automation models by maintaining the ML decision-trees and up-to-date digital knowledge base in ever-changing and evolving network operations. It’s an interesting change of an old paradigm. Though it won’t happen overnight, it will be an increasing shift that we expect to see develop over the next several years.

Low Barriers to Entry

The initial perception about integrating automation solutions is that the cost, time and resources will be high. In reality, it depends on the solution and the business. The technologies leveraged for automation solutions are typically not massive systems that need to be developed. In many cases, the investment to get started with automation is minimal. We have put a lot of focus into making these aspects very simple in our own solutions: how to create the models, how to train them and how to maintain and add further value to the automation functions over time. As a result, the important issue for operators to focus on when beginning to explore automation solutions is not the cost, but the value provided to the customer.

With automation systems, you can demonstrate the ROI very quickly, often within just a few months. While there is complexity, if there is a solid underlying system to quickly create automations using the engineering workflows, you’ll see benefits immediately. It’s an interesting model, and since the ROI is there, it’s quite an easy investment decision from an operator’s point of view.

Long Term Value of Operational Proactiveness

Roughly 30% of overall customer care issues, including churn, are related to something technical, and those issues end up landing in the hands of the operations and engineering people. They spend a significant amount of time going through issues, looking at several different tools to correlate the root cause of the problem, and then determining how to fix and communicate it back to the customer. This takes a long time and is a heavy burden on the engineering team.

We set out to automate technical customer care. From day one, our slogan has been “Operations Made Simple.” Much of our efforts in automating technical customer care are now moving towards not only highly accurate and fast “reactive resolutions”, but also developing into proactive resolutions and actions. Additionally, we solve for automating network engineering and optimization tasks. Our Network Advisor product in this field recently won a Fierce Innovation Award from Fierce Wireless.

With automated technical customer care, you start by leveraging the engineering expertise on how to determine the right route for customer issues and build a model. Over time, you continuously add more new information to this dynamic system in order to get more granularity and more accuracy to pinpoint and resolve issues. Through AI techniques, the system continues to adapt and scale to increasing complexities and changes in the network. This is how we can add new value to an existing process with automation, and the savings in time and costs are quite significant and most importantly, the customers are kept happy.

The future of automated customer care resolution will go beyond improvements to existing workflows, and into the next phase where instead of reacting to a customer complaint, the system proactively identifies a potential issue, correlates the root cause, and notifies the engineer to resolve it before the customer is affected or experiences it. In some cases, the system will even auto resolve the issue without an engineer getting involved. As a result – the issue is pre-empted, and customers don’t churn or call to complain. This is the holy grail of customer care that artificial intelligence, machine learning and automation are challenged to solve.

What’s Next?

The next five years are too far out to be written, but we expect the solid foundations in automating key use cases like technical customer care resolution will set the cadence for what’s to come. While we see operators being the most aggressive pursuers of automation technology in the wireless industry due to the increasing complexity of networks, IoT is also emerging as another key service to be managed with automation in the future. The possibilities for automation are endless, and we’ll be working on solving the next big challenges in the wireless industry.