Use cases for AI in Telco: An era of mainstream adoption

Report Overview

Use cases for AI in Telco: An era of mainstream adoption

Already a Subscriber? Login here to view Full Report

Use cases for AI in Telco: An era of mainstream adoption
Many tier 1 telcos have begun to implement AI initiatives to build and operate their network, sell more efficiently and improve the customer support experience.
Based on our industry interviews, we expect increasing competitive intensity in the telco market to drive increased adoption of AI beyond the largest incumbents and into RSPs
Many large telcos are facing a myriad of headwinds: declining bandwidth prices, a more agile and less loyal consumer base and significant margin transfer to content providers. Many incumbents also face the issue of legacy systems that do not have the capability to deliver services to the next generation of telecommunication networks and an increasingly demanding customer base. Over the past 10 years, global companies have seen an average decline in margin from 35% to 33%, with net income dropping by 25%[1].
One of the biggest constraints facing telco companies is their ability to process and interpret data and use business intelligence to quickly respond to customer and market demands. Network, customer and market data could represent tens of millions of data points in a given day. While advances in Big Data provide the processing capabilities to store and process the data, analysis relies on manual interpretation from the network, marketing, and sales teams.
Over the past few years, artificial intelligence (AI) systems have been adopted to derive insights from these masses of data, and as the development of AI progressed and algorithms and techniques improved, the scope of AI implementation has expanded.
Deloitte research recently showed that telcos, in comparison to other industries, have invested more in AI and seen greater returns for their investment. The high ROI for AI projects therefore comes with the promise to reverse the downward trend in margin and income.
The first step is to identify key use cases that can add value. This paper aims to describe use cases that have been implemented in industry, to determine the impact of these efforts and therefore highlight key areas to focus on where building out AI capacity.


Key takeaways

Artificial Intelligence is hitting critical mass in Telco

  • Global telco service, software and hardware providers are building out capabilities

Network, customer service, and sales and retention are the most common applications of AI in telco

  • Moving from business intelligence to machine learning
  1. Network design and operational use cases
  2. Customer experience AI use cases
  3. AI for Sales and Churn Prevention use cases

Factors driving AI adoption within an organisation

  • Providing clear business cases with cost estimates and benefit targets
  • Building a clear team structure with defined funding sources
  • Alignment to corporate and organisational strategy

Taking a fast-follow approach by leveraging global service providers

  • How ready is your organisation to make AI a leading part of its strategic agenda?
  1. Are you using AI to support your corporate strategy?
  2. Are you extracting deep, actionable insights from your customer data?
  3. Are you struggling to efficiently scale your operational capacity?
  4. Are you maximising your return on infrastructure deployment?
  5. Have you chosen the right team structure to ensure wide scale adoption of AI use cases?

Conclusion: Adoption of AI is already table stakes

Appendix 1: Primer on ‘artificial intelligence’

What do we mean by ‘Artificial Intelligence’?

  • The history of AI
  • Narrow AI versus general AI

AI: Sorting the hype from the reality

AI Glossary

List of charts/tables

Figure 1. ROI vs investment for various industries

Figure 2. Example mobile heatmap of the received customer network quality

Figure 3. Conceptual diagram for the self-healing network

Figure 4. Example of a CVC upload and download traffic

Figure 5. Examples of Telstra and Optus chat bots

Figure 6. Conceptual diagram for building propensity models

Figure 7. Timeline of AI reaching human-equivalent performance across narrow tasks

Figure 8. Annual AI-related published per year

Figure 9. Comparative number of research papers published on deep learning, by country

Figure 10. AI expert predictions on timeline for matching human performance

Figure 11. Perceptions of emerging technology in banking among CIOs