• Topics
  • Research & Analysis
  • Features & Opinion
  • Webinars & Podcasts
  • Videos
  • Dtw

Delivering 1:1 real-time marketing with Generative AI

The GenAI hyper-personalized customer experience Catalyst has created a highly reusable automated marketing framework for CSPs, capable of providing bespoke customer communications and product criteria

Ryan Andrew
30 Oct 2023
Delivering 1:1 real-time marketing with Generative AI

Delivering 1:1 real-time marketing with Generative AI

Commercial context

Providing the highest-quality product requires a keen understanding of customer needs. When feasible, this involves creating bespoke offerings that meet them down to the precise detail, when and where the customer wants – but customers’ unique preferences generally make doing so at scale impracticable. Many communications service providers (CSPs) don’t have the resources to assess and respond to the great variety of desires of all their customers, and attempts to understand individual customer preferences in depth typically enjoy limited success, and can in fact themselves lead to frustration and customer churn.

The emergence of predictive analytics along with generative AI however has made hyper-personalized approaches at scale, a realistic prospect. By joining the dots across the variety of channels that customers use, and assembling a real-time representation of their desires as well as any frustrations (such as service interruptions), such approaches can provide highly tailored communications which are significantly more relevant, engaging and empathetic. Hyper-personalization therefore takes a huge step towards provision of truly bespoke products and services, and stronger customer loyalty.

The solution

The GenAI hyper-personalized customer experience Catalyst was designed to establish a highly reusable and automated marketing framework for CSPs. The Catalyst proposes a radical shift to a cloud-native, AI-driven framework for digital engagement. Its purpose is to adapt continuously to customer needs by providing unique recommendations and resolving user problems – and, crucially, to produce a way of doing so which is zero-touch, on-demand and can be executed at scale. Success here means CSPs can fundamentally improve their operational marketing models and decision-making speed.

There are three core pillars to the technological logic of the solution: complex event processing (for ingestion of data from multiple sources, such as customer data, network events and clickstream events - and their transformation into actionable insights), predictive ML modelling (combining real-time events to calculate customer churn propensity score), and genAI. The solution’s architecture is based on the zero-touch interoperability model of Open Digital Architecture (ODA) – predicated on virtualized infrastructure, cloud-native IT components and AI-enabled real-time operations. Using Google Cloud’s Vertex AI and PaLM2 fueled Generative AI, Pega’s Real-Time Customer Decision Hub, and Accenture Customer Data Platform (CDA), the project team has curated a set of ODA-compliant functional components with a ‘write once, deploy anywhere’ model.

The standardized components are connected via open APIs, and the data models used are both SID and AI-optimized. By using ODA economics, the solution allows CSPs the flexibility to invest in differentiating capabilities. Open APIs are used to connect ODA functional blocks of Party, Core Commerce, Production, Engagement and Intelligence Management. TMF 629 and 717 are used for party management, and TMF 622 is used to place or change orders to downstream order management systems when an offer is accepted or retained. TMF 680 is used to deliver omni-channel recommendations in real-time – so genAI can curate parts of the recommendation and syndicate it, whether it’s an SMS, email, eChat prompt, IVR playback, push notification or button on a website or app.

With five leading CSPs participating in the Catalyst (Bell Canada, Turknet, BT Group, Telecom Italia and Deutsche Telekom), it’s a clear sign of the industry’s determination to meet rapidly changing customer expectations. Collectively, they have contributed a wealth of customer and network insights to create an innovation which is reusable across markets. By blending these insights with customer data into Accenture’s CDA, Pega Customer Decision Hub is enhanced with real-time machine learning model scoring through Google Cloud’s Vertex AI and real-time insights to deliver relevant on-the-fly decision-making personalization at the moment of engagement with the customer.

Adding gen AI capabilities from Google Cloud enables curated marketing campaigns, with tailored, hyper-personalized text and image generation. To ensure the real-time operation of the solution, the data and Pega’s decisioning need to be deployed in very close proximity to Vertex AI, on top of the highly secure and performant Google Cloud. As a result, this very low-latency and powerful real-time AI engine creates differentiated outcomes that are truly pioneering.

Application and wider value

According to Neel Mehta, Bell Canada, Director IT Delivery “this Catalyst is truly groundbreaking. The churn use case benefits are obvious and we foresee wide adoption of this architecture to create a highly differentiated Bell Canada-specific execution, despite standardization. We see the potential to create key differentiators in enhancing customer engagement, increasing conversion rates, improved customer satisfaction, optimized marketing spend, and the ability to substantially reduce average handling time of agents at our call centers. Apart from pure marketing use cases, we see great potential in the use of this pattern in creating new opportunities in the industry such as use in network operations, auto-healing, buy-flow automation, agent workflows and more.”

By creating the means for digital platforms to share their diverse perspectives, ideas and knowledge, AI and machine learning can provide hyper-personalization of content and services that can be tailored to individual preferences. The more that is invested in this system, the greater the complexity of preferences it is able to undertake. While the principal use case is designed to reduce churn, its applications are limited only by the human imagination. When deployed at scale, people will be able to see their preferences realized with fewer and fewer onuses.

As CSPs across the globe seek to use AI as a means to demonstrate innovation and secure new revenue streams, this Catalyst shows the industry how. By providing a working example of genAI capable of understanding the customer’s frustrations and desires in real-time, CSPs are in a prime position to provide anyone and everyone with services they truly want and need, and help build a more connected and inclusive society in the process.

Catalyst: GenAI hyper-personalized customer experience