Make it happen – Considerations for CSPs when adopting AI
Done correctly, AI can redefine future customer experience for CSPs, instead of them following the lead of digital competitors.
12 Oct 2018
Make it happen – Considerations for CSPs when adopting AI
This is an excerpt from our paper AI & customer experience: Emerging best practices. Keen to learn more about how AI can improve customer journeys and experiences? Visit this page for the full report.
Success is not just about understanding AI, but working across an organization to create use cases and identify points of commonality. A leader in the CSP organization must be empowered to do this likely a CIO or chief data officer. AI only works if it is thought of holistically across the entire company.
To obtain budget, many AI initiatives are developed for specific use cases. This has many benefits as investment and impact can be monitored and assessed. However, a pure use-case approach risks creating AI silos, meaning the same lessons are learned over and over again but cannot be implementated across traditional business silos which would produce some of the greatest benefits. In time, some form of centralized development and coordination will become essential, but it pays to avoid being too prescriptive about the exact shape and extent of that centralization too soon.
Many CSPs have some years of experience with analytics under their belts now, particularly around their customer base and value management. As much as two-thirds of the opportunities to use AI are in improving the performance of existing analytics use cases. Yet one of the biggest problems CSPs have is that their data are often siloed, so insights from one division aren’t shared with other areas of the company.
An example is how focused CSPs are on their channels so that they miss seeing the whole customer. Operators need to shift from a channel first mentality to a customer-first ethos when it comes to analytics and AI.
Linking data across customer segments and channels, rather than allowing the data to languish in silos, is especially important to create value.
If it’s a broken process for humans, it will be broken for machines. CSPs should not introduce AI as an overlay and expect miracles; rather they need to identify problems, and redesign customer processes to make them digital from end-to-end and suited to automation. Only then can they understand where and how AI can best deliver value. It is importants to understand that AI cannot solve all issues, although there is more synergy between AI and analytics than differences.
AI solutions should be integrated into existing IT, customer relationship management, operating software and business software systems using RESTful APIs, such as TM Forum’s Open APIs. CSPs must be easy to do business with if they are to work with the best AI technology suppliers. This includes having standardized approaches to data collection and management to enable the suppliers to find a common approach to plugging CSPs’ data into their platforms.
Due to the problem of siloes, one of the first things AI companies tend to do is spend time with each department to determine their maturity in using data intelligently to improve business processes, and interactions and data sharing between groups to drive common business goals.
Empower a leader to work with business units
Success is not just about understanding AI, but working across an organization to create use cases and identify points of commonality. A leader in the CSP organization must be empowered to do this likely a CIO or chief data officer. AI only works if it is thought of holistically across the entire company.
Amazon is an outstanding example of having a holistic AI strategy; each business unit is encouraged to leverage ML and AI. Successful innovation in one division is applied in other areas. Breaking down the silos encourages innovation across the board.
Coordinate AI initiatives
To obtain budget, many AI initiatives are developed for specific use cases. This has many benefits as investment and impact can be monitored and assessed. However, a pure use-case approach risks creating AI silos, meaning the same lessons are learned over and over again but cannot be implementated across traditional business silos which would produce some of the greatest benefits. In time, some form of centralized development and coordination will become essential, but it pays to avoid being too prescriptive about the exact shape and extent of that centralization too soon.
Integrate AI with analytics for efficiency and synergy
Many CSPs have some years of experience with analytics under their belts now, particularly around their customer base and value management. As much as two-thirds of the opportunities to use AI are in improving the performance of existing analytics use cases. Yet one of the biggest problems CSPs have is that their data are often siloed, so insights from one division aren’t shared with other areas of the company.
An example is how focused CSPs are on their channels so that they miss seeing the whole customer. Operators need to shift from a channel first mentality to a customer-first ethos when it comes to analytics and AI.
CSPs will also have to adopt and implement strategies that enable them to collect and integrate data at scale. Even with large datasets, they will have to guard against “overfitting,” where a model too tightly matches the “noisy” or random features of the training set, resulting in a corresponding lack of accuracy in future performance, and against “underfitting,” where the model fails to capture all of the relevant features and so has limited usefulness.
Linking data across customer segments and channels, rather than allowing the data to languish in silos, is especially important to create value.
Automation without AI is only half the solution
AI automates processes to create scalable insights from large amounts of data. Technically this is true, but it is not magic dust that can make a process work better just by adding intelligent automation.
If it’s a broken process for humans, it will be broken for machines. CSPs should not introduce AI as an overlay and expect miracles; rather they need to identify problems, and redesign customer processes to make them digital from end-to-end and suited to automation. Only then can they understand where and how AI can best deliver value. It is importants to understand that AI cannot solve all issues, although there is more synergy between AI and analytics than differences.
Integrate AI with Open APIs
AI solutions should be integrated into existing IT, customer relationship management, operating software and business software systems using RESTful APIs, such as TM Forum’s Open APIs. CSPs must be easy to do business with if they are to work with the best AI technology suppliers. This includes having standardized approaches to data collection and management to enable the suppliers to find a common approach to plugging CSPs’ data into their platforms.
Due to the problem of siloes, one of the first things AI companies tend to do is spend time with each department to determine their maturity in using data intelligently to improve business processes, and interactions and data sharing between groups to drive common business goals.