Automation via AI/ML/RPA has its unique and influential role in addressing challenges and opportunities that Telecom providers face today. Per a recent survey, 94% of business leaders surveyed agree that Automation is critical to success over the next five years and can unlock $700 billion in value for the telecom industry. At the same time, while Automation adoption has increased to 50%, only 8% of the companies have achieved more than 20% increase in profits.

Telcos have amassed a vast amount of customer data over the years, opening many opportunities to understand and predict customer behavior to improve customer sales & service operations. Similarly, telco networks generate a lot of data that can be used to enhance network and service operations.

To keep up with their objectives and run effectively, Telcos have felt the need to automate rapidly. Hence, the role of RPA, AI & ML in Automation is often scrutinized. And do the 3-technologies complement or compete? To ensure the success of the Telco’s, Industry needs a framework focussing on Machine Learning (ML) driven Robotic Process Automation (RPA) backed by Artificial Intelligence (AI) to think.

Telco needs to create an Ecosystem collaboratively and focus on four key aspects to start their journey.

1. Skills

According to the Gartner survey, the top barrier to Automation adoption is a lack of skills. 56% of leaders acknowledge that Automation will change the skills needed to accomplish Automation jobs. Automation has seen massive growth in the last five years, but the talent pool is still limited. Since, digital native companies were the pioneers in this field, most of the talent pool is absorbed by these companies creating an even greater shortage of talent with Automation skillsets.

Telcos need to invest in retraining and up-skill their existing employees as well as hiring talent from outside. Automation initiatives not just need data science and AI/ML/RPA experts but also business and domain experts who can help give meaning to data. For example, one of our clients built a cross-functional team of business and IT professionals to jointly work on an Automation initiative that delivers the desired outcomes by making all team members (business & IT) equally accountable.

2. Trust

According to the 2021 Forrester survey, respondents claim trust (accountability, transparency, security) is a big challenge in adopting AI/ML. Decision-makers are unable to maintain oversight and governance of machine decisions and actions and are concerned about unintended, potentially harmful, and unethical outcomes. In addition, people’s inability to understand how AI/ML produced its result is a cause for alarm for many, as human beings cannot trust something they cannot explain.

AI/ML algorithms take time to learn data, and it does not start working immediately. As an example, it can predict failures in the network, but at the same time there might be a lot of false positives in the beginning. The operations team will not trust AI/ML if it’s only accurate 50% of the time. It takes quality data and multiple trials to improve the accuracy and to get to an acceptable level of accuracy.

To overcome trust issues, telcos need to start slow and still heavily rely on people to make decisions based on the insights gained from AI/ML. However, once the trust begins to build up, Automation can effectively automate the process instead of a person first validating the decision.

3. Data

Siloed data, and poor data quality, are some of the significant challenges that telcos need to focus on to harness the full potential of Automation. Most of the telcos have accumulated multiple data warehouses through acquisitions. As a result, there are multiple data lakes, data sitting over at multiple places, and inconsistent business definitions create data integrity challenges. Additionally, data might be sitting at private, public or hybrid cloud environments adding to the complexity.

Data is the core underlying building block that significantly impacts Automation outcomes. AI/ML systems need data to learn, identify patterns and produce meaningful outcomes. Therefore, it is vital for telcos to invest in comprehensive enterprise data architecture and embrace the importance of data early on in their Automation journey. For example, one of our clients is building predictive analytics capability on Google Cloud by ingesting all customer data collected across various touchpoints, including network sourced internally and from third parties.

4. Operating Model

Siloed approach to building point solutions is nothing new for enterprises; unfortunately, the same trend has continued with Automation implementations. As an example, the Marketing team might have their own Automation solution to drive the effectiveness of marketing campaigns while the customer service team is focused on improving their operations by leveraging a different implementation. While providing some outcome, both solutions fail to recognize the fundamental promise of Automation technology to stitch together a holistic view of a customer to predict future behavior correctly.

One of the ways our client has solved this challenge is by setting up a COE. The COE includes a business unit lead, which identifies and prioritizes opportunities. Federated teams across all business functions then leverage the COE. Telcos need to rethink their Automation operating model focusing on AI/ML/RPA capabilities that can be leveraged consistently across business functions.

In conclusion, telcos need to find ways to overcome the above key challenges associated with Automation adoption. First, build a talent pool and required skillset. Starting small and leveraging data to make decisions while still leaving the final decision in the hands of the decision-makers instead of AI/ML is an excellent way to start addressing trust issues. Identify comprehensive data needs for Automation initiatives sourced internally and externally. Finally, re-imagine the Automation operating model to truly benefit from the latest technological advancements.

• Global AI In Telecommunication Market Size, Status and Forecast 2021-2027 (valuates.com)

• 141 Myth-Busting Statistics on Artificial Intelligence (AI) [2023] (aimultiple.com)

• us-ai-institute-state-of-ai-fifth-edition.pdf (deloitte.com)

• The state of AI in 2022–and a half decade in review | McKinsey

• Telco Industry Technology Transformation | Accenture

• Telecoms have unique challenges in adopting AI | TechRepublic

• How to benefit from AI in telecommunications – N-iX

• Adoption of AI advances, but foundational barriers remain | McKinsey

• What’s holding back telecom adoption of artificial intelligence? (accedian.com)

• CenturyLink’s No Sweat Approach to AI | Light Reading

• 3 Barriers to AI Adoption (gartner.com)

• Bridge The Trust Gap Between AI Technology And Impact (forrester.com)