Why Telecom Needs AI, Not Just Automation


The telecommunication service provider (mobile, fixed, cable, and broadband) industry is one of the biggest businesses in the world. It also has historically been a capital-intensive industry with high fixed costs, which has put pressure on telecom operators to control their variable costs, particularly human capital. Tom Nolle of CIMI Corporation estimates that many telecom operators crossed the point where revenue per bit is lower than cost per bit in 2017. Telecom operators are straining under digital transformation, which Bill Schmarzo, chief technology officer (CTO) of Dell EMC Services defines as: “The coupling of granular, real-time data (e.g., smartphones, connected devices, smart appliances, wearables, mobile commerce, video surveillance) with modern technologies (e.g., cloud native apps, big data architectures, hyper-converged technologies, artificial intelligence, blockchain) to enhance products, processes, and business-decision making with customer, product and operational insights.” On top of all that, telecom operators must solve how to profitably manage and operate the dizzyingly complex next-generation 5G/IoT networks.

It is an industry ripe for AI-driven solutions, with their promise of lowering costs and boosting efficiencies. Many telecom operators have begun to experiment and deploy AI-driven solutions in both customer-facing and internal organizations. In a new report, Tractica has identified seven key telecom AI use cases: network operations monitoring and management, predictive maintenance, fraud mitigation, cybersecurity, customer service and marketing virtual digital assistants (VDAs), intelligent customer relationship management (CRM) systems, and improving customer experience management (CEM).

Many experts believe a significant amount of telecom automation can be achieved through deterministic software solutions, with little impact from AI solutions. While deterministic solutions will play an important role, Tractica believes AI-driven solutions will be even more critical if telecom service providers are to survive.

Managing a Dynamic Network

Networking software player Aria Networks has built an AI-based network optimization solution that, today, goes beyond prompting network engineers. Customers include Verizon, Telus, Level 3, and other Tier One operators.

“If the goal is automation (and perhaps even autonomy), it’s an incomplete strategy to focus only on identifying patterns,” wrote Robert Curran, Aria’s chief marketing officer (CMO), in an email to Tractica. “In a network-centric business like telecom, understanding how decisions are made about how, and when to change the network are the real heart of the operation. That’s where automation efforts should be focused.”

Curran said the trajectory for telecom is to get to much more dynamic networks, with more software defined and controlled. “When you can make changes in software, then you can go faster, so you need to make decisions more quickly,” said Curran, “Network design becomes a more complex problem. 5G and IOT will complicate this further. Some of these networks are so big, humans can’t efficiently manage them.”

Curran gave some examples of where Aria’s automated network management solution is solving network issues:

One of our customers has to send out [internet protocol television] (IPTV) signal across the network. For protection purposes, it has to be sent from two different studios. The problem is how do you compute that so the service meshes and synchs? You need to do it so you can optimize and get more traffic on the networks. It’s a time scale complexity issue. Using conventional methods, you would have to pre-test it a week before. Another example of how the solution works might be maximizing cost efficiencies. A video service sending content to a location, the cost of that from a data center, power and transport costs vary, so if you could move that to a more cost-friendly center, that would be good. A third example would be ensuring revenue streams. Let’s say you are televising a Formula 1 event and want to sell click through. If that signal degrades, you lose revenue, because you can’t sell that click through.

Improving Network Performance

Since 2015, Nokia has made investments to position itself as a contender in AI-fueled network management and operations and it appears to be paying off as telecom operators migrate to 5G. In February 2016, Nokia launched its AVA platform, a cloud-based network management solution leveraging machine learning (ML). According to the press release, AVA features “A centralized big data collection and storage processing layer, supported by machine learning methodology, gathers data from sensors in network elements and OSS.” Nokia’s commercial deployment of AI-driven solutions grew from there. In October 2017, the company announced the commercial launch of an AI-driven Analytics Service to run on top of AVA to better manage telecom networks. Nokia describes some use cases within network management where Analytics Services can help:

  • Spectral Performance Management: This enables granular capacity planning.
  • Cell Site Degradation Prediction: Nokia claims it can predict service degradations on cell sites up to 7 days in advance using ML.
  • Similar Ticket Recognition: This uses DL to find patterns in unstructured data, then Nokia’s virtual digital assistant MIKA guides engineers to suggested best solutions for network issues. Nokia claims this leads to a 20% to 40% improvement in first-time resolution.

In February 2018, Nokia unveiled its 5G network architecture called Future X at Mobile World Congress. The architecture leverages AI in the core and radio access networks (RANs) to “improve both performance and operations.” Like larger rival Huawei, Nokia has announced its 5G beamforming antennas will be assisted by ML to extend cell range and manage network capacity per antenna.

Automating Service Delivery

In telecommunications, CEM refers to managing the experience and quality of service. While there are a number of overlaps in functionality with traditional and “intelligent” CRM systems, CEM is used in telecommunications to support various elements of customer experience that most CRM systems are not equipped to handle. Examples include auto-adjusting network parameters, service quality detection, website quality detection, and addressing network performance or security needs in real time.

Mobile and network service providers are now leveraging AI in a number of ways that both enhance customer experience and help automate quality of service. Chatbots and virtual agents can use ML and natural language processing (NLP) to handle support interactions via a short messaging service (SMS) or other messenger platforms and make necessary changes or updates. Network and device data can be used to predict and preemptively execute provisioning or other automation to optimize reliability. Real-time rating, charging, and meditation capabilities can streamline billing processes. Ongoing qualitative and quantitative customer interactions, requests, complaints, service logs, and cross-channel portals can be analyzed using ML, NLP, and deep learning (DL) to detect trends or performance issues across demographics, devices, time, or location. Having integrations across CRM, operations tools, call center solutions, social analytics, etc., enables AI to help CEM systems convert interactions into insights across the entire customer and device life cycles.

Startup DeviceBits is leveraging AI expertise to provide telecom operators with a solution for both the predictive maintenance and the improve customer experience management use cases by delivering predictive, self-help-based customer support. Chief Executive Officer (CEO) JC Ramey explained:

We have found with our telecom customers that the first thing they want to do is empower the customer, though I think that is less them driving and more about real consumer demand for self-help. What happens when the problem the end user is having is related to a network problem? We have been working with bigger cable and mobile operators to see what is going on in the network, with the goal of incorporating network visibility into self-help. That not only helps the telecom operator to avoid the call but how to have the end user fix it themselves. So when you come home and turn on the TV, there is a message that says, “Hey, we know you have an issue, turn off your router, etc.,” so they get a step by step instruction on what to do.

Ramey said in terms of where things stand, DeviceBits is making some strides with self-diagnostics with consumer electronics, adding:

We are using SMS for alerting people to take some action. We have good visibility into the home or devices, now we need it to turn it into action. At the individual or personalization level we are getting there, we are collecting a lot of data, 170 million telecom specific questions a year are being fed into the system, so we have good problem to resolution mapping. What we need now is this continuous overlay and like anything, to understand what are the gaps in our information or telemetry data to accurately predict what’s going on. With this information we can turn a lot of academic AI into practical AI for our customers.

The key area where AI plays a role in the solution is processing the telemetry data, predicting the issue, the cost associated with the issue, and the timeline for when the issue will occur.

AI-driven solutions for telecom operators will not occur overnight. Due to the massive complexity of telecom networks and other factors, AI-driven solutions will become more common within the telecom industry over time, particularly as 5G networks are deployed. AI-driven solutions will be necessary as service providers move toward the utopian autonomous network.

Comments are closed.