5G Chargers – using data and AI to tailor services in real time, at scale

The ambition of this Catalyst project is to create a modular, end-to-end architecture for dynamic catalog and 5G charging to offer personalized services dynamically, in real time. It shifts the product catalog to the heart of operations, enabled by AI across the portfolio.

The Catalyst, 5G Chargers, demonstrates how CSPs can start to contextually and dynamically personalize the customer experience, in real time and at scale, by implementing an open, componentized, modular IT architecture, driven from the product catalog, as opposed to billing and charging being at the core of operations.

The proof of concepts champions, Orange, Telenor and Vodafone, want to find solutions to several challenges. They need more business flexibility to create more services, and to personalize services for different customers while simplifying their business. In short, the operators want to do more with less.

The champions are supported in the Catalyst by participants Oracle, SigScale and Whale Cloud. The team has built a product catalog-driven platform (as shown below) using TM Forum’s Open Digital Architecture and Open APIs. The platform controls the commercial-use rules.

The team also drew on work from last year’s award-winning BOS Catalyst, which created a componentized architecture using the TM Forum Canvas, Component and other ODA elements. To demonstrate the possibilities of the AI and catalog-driven platform, this year the team demonstrates three scenarios.

Dynamic offer creation for B2B2X

If a CSP partners with a media or gaming company whose offers involve augmented or virtual reality (AR/VR), for example, the CSP can commit to ensuring that their partners’ customers get the best quality of service when consuming the partners’ services, such as by guaranteeing no buffering. It might be that the partners are willing to pay a higher percentage of revenue share for that experience guarantee.

In this instance, the team has set up a flow whereby when a partner joins the CSP’s ecosystem, the portal analyses the products the partner is onboarding, based on a number of parameters. They include the necessary quality of experience, and draw on the CSP’s own experience of delivering services for partners with similar characteristics combined with AI to gain insights to recommend the type of network experience the partner and their end users need.

That could be best effort, or it might need a network slice already provisioned by the CSP with new parameters added, or the CSP might need to spin up a new slice to meet the criteria.

There is a product catalog element to onboard and deliver partners’ services as well as a charging component to determine if an additional fee is associated with the quality of experience. Whale Cloud provides AI capabilities, product catalog, sales portal, product order management, customer information, party identification and authentication, API Gateway in the first two cases .

Creating dynamic promotions

The B2C second use case feeds off the first: At some point, CSPs will have multiple partners in their ecosystem requiring a product catalog that's flexible and modular. This means that CSPs can right-size products for customers wanting a new service through the AI-powered, dynamic catalog – as opposed to one with fixed priced products. This could include speed, capacity, price and other parameters, depending on what the customer wants to do.

Context as well as dynamism is critical to understanding customers’ behavior and patterns in that behavior, so the AI analyzes the product catalog itself, and the different components within the product catalog. The aim is to be able to tune and configure the qualities of each component in the catalog to create the best and most appropriate offer at any time, including for multiple services simultaneously in some instances.

Dynamic rate charging for roaming

Today CSPs often offer best efforts services, charged for by the amount of data consumed, regardless of whether the customer was watching a movie, browsing online or checking email. When roaming, the same principles apply today – it’s the amount of data rather than how it’s consumed that CSPs charges for.

Oracle provides the AI-driven rating component in this use case, working with SigScale’s charging function, to offer more personalized, dynamic pricing and charging based on customers’ behavior. In this scenario a user from Europe who is an avid gamer travels to Chile. She also likes video services. Instead of offering her a euro-per-Megabyte deal, the CSP can offer a special rate for gaming, because the operator wants to keep her on the network and engaged, and could perhaps offer some discount for video streaming too.

Using AI, the CSP is able to better understand the user and her behavior for example from her usage patterns, account standing (including customer lifetime value), and propensity to churn. The AI-driven rate selector takes these and other factors, such as the CSP’s preferred margin on the service, into account to before selecting and sending a custom rate to the charging function. This process is completed for each service the user consumes while roaming in Chile.

The idea is to be able to approach every customer in this way, based on their wants and needs. By taking in and analyzing all these pieces, the aim is to build customers’ loyalty, keep them on the network and encourage them to take up new services.

Next steps

In defining the scope of this project, the team realized that TM Forum and 3GPP are closely aligned in terms of their objectives, but there is little overlap in some areas. For example, 3GPP standards define the rating function, but TM Forum lacks Open APIs for rating. The team intends to contribute this to the Forum’s suite of APIs and to facilitate additional collaboration between the Forum and 3GPP to address any gaps between their standards.

Visit the Catalyst space for more information from the team.