Summary

This report analyses the use of AI-driven zero-touch1 automation in a telecoms environment, within which it can play a central role in automating complex end-to-end processes. We examine the technology required for zero-touch automation, discuss how this technology will be used and explain what zero-touch will mean for CSPs and vendors

ページTOPに戻る

Table of Contents

About this report
Executive summary
Executive summary
Analysis and recommendations
Challenge: the complex and dynamic nature of end-to-end processes in the telecoms sector has made ‘zero-touch’ automation challenging
Solution: AI provides a solution for more-complex automation issues and helps to achieve zero-touch end-to-end processing
Solution: AI enables complex processes to become candidates for ZTA by reducing the cost of modelling and by delivering insights and algorithms
Solution: AI can be embedded during the development of applications and processes to replace hard coding with insights on which to act

Recommendations

How is AI applied to zero-touch processes?

There are four main ways that AI enables zero-touch automation to replace manual intervention within a given process

Key component-level integration development and AI infrastructure are being used to build the AI ecosystems that are needed to deliver zero-touch automation

Zero-touch automation requires an abundance of good-quality data to initially model each use case and to be able to continue learning by monitoring outcomes

Why is AI needed to provide zero-touch automation?

AI helps to address complex processes that have numerous decision points and multiple data inputs; it also helps to prevent a significant number of rules variations

Zero-touch automation has limits unless AI technology is applied to it; AI technology has matured sufficiently to deliver its full capabilities in a cost-effective manner

Replicating human intelligence and beyond requires an understanding of the real world in order to reproduce and build on human intelligence

Where is AI being used in zero-touch automation?

Automated monitoring processes for complex data sets are possible when using ML to find exceptions or anomalies within a standard model

Diagnostics and RCA can help to identify problems with the underlying data in order to predict and prevent (or reduce) undesirable outcomes

Zero-touch automation makes more-precise sales and marketing possible, and can help provide market insights at a micro-segmentation level to each user/subscriber

Customer engagement processes have traditionally involved substantial manual intervention; shifting to zero-touch automation requires real-world understanding

Appendix

Case study: Elisa

Case study: Verizon

Case study: Orange

Case study: Deutsche Telekom

About the author and Analysys Mason

About the author Analysys Mason’s consulting and research are uniquely positioned Research from Analysys Mason Consulting from Analysys Mason

List of figures: 

Figure 1: AI-based approaches to zero-touch automation build on other approaches and address more-complex and dynamic issues
Figure 2: AI use cases tend to focus on complex automation issues that are expensive or not possible to do manually
Figure 3: Scope of zero-touch automation
Figure 4: Life cycle for modelling data and generating
insights
Figure 5: Application development is now linked with calls
to AI insights and algorithms
Figure 6: The four levels of creating a zero-touch process

Figure 7: AI stack options built into development environments

Figure 8: AI workflow for development and deployment

Figure 9: Logic fabric can become highly complex even with a simple four-step process
Figure 10: AI provides an increasing level of support for more-complex processes, building on earlier technology

Figure 11: The skills that are needed to interpret raw data help ZTA solutions to fully understand different data.

Figure 12: Using AI to monitor processes is straightforward
Figure 13: Diagnostics of cause and effect are inherent within AI-based modelling.

Figure 14: The ways in which AI can be used in marketing and sales processes for zero-touch

Figure 15: How AI works for customer interaction and with chatbots