Login
Register
This Moonshot Catalyst shows how an AI-native, multi-agent architecture can transform customer service, sales and operations by interpreting intent, orchestrating specialist agents and resolving customer issues faster, more accurately and more proactively.
Agentic AI gives customer care a proactive edge
As communications service providers move towards AI-native operations, customer experience is becoming one of the clearest tests of whether automation can deliver measurable business value. Customers expect faster answers, more personalized support and fewer repeat interactions, while service providers need to reduce manual effort, improve resolution rates and create new revenue opportunities beyond connectivity.
The Catalyst, GenAI proactive customer experience: AI realm agentic driven architecture - Phase II, addresses this challenge by applying an AI-native, multi-agent architecture to customer service, sales and operational automation. Developed as part of TM Forum’s Moonshot Catalysts under the Composable IT and Ecosystem Mission challenge, the project explores how intent-driven automation can support end-to-end customer journeys, from AI-powered contact centers and assisted sales to proactive care and autonomous service orchestration.
The team is targeting practical gains across customer experience and operational efficiency, including a 30% increase in Net Promoter Score, a 30% improvement in first contact resolution and a 30% reduction in average handling time. It also aims to automate more than 85% of standard customer queries through AI self-service bots, improve outbound call answer rates by 20%, accelerate claim closure by 20% and achieve more than 90% intent recognition accuracy.
At the center of the Catalyst is a dynamic orchestration layer that interprets user intent and routes requests to specialized AI agents, each responsible for a specific domain or task. This replaces monolithic workflows with modular, composable services that can be dynamically assembled to manage complex customer and operational scenarios.
The architecture combines large language models, enterprise knowledge systems and ontology-driven reasoning to deliver responses that are context-aware, precise and aligned with business rules. Retrieval-augmented generation grounds interactions in trusted data sources, while intent-based decomposition breaks complex requests into smaller tasks that can be handled autonomously by different agents.
These agents collaborate through coordinated orchestration, enabling end-to-end automation across customer care and enterprise operations. The Catalyst also incorporates monitoring, evaluation and feedback mechanisms to track agent performance and improve outcomes over time. This creates a continuous improvement loop in which the system can adapt to new scenarios, refine responses and increase automation quality.
The Catalyst demonstrates how traditional BSS and OSS environments can evolve into intelligent, AI-native platforms. By enabling systems to understand intent, reason over enterprise knowledge and coordinate action across distributed components, the project supports a shift away from static process automation towards adaptive service orchestration.
For customer-facing teams, this could mean faster and more accurate handling of routine inquiries, more proactive service interventions and more effective sales journeys. For operational teams, it offers a way to reduce manual effort, monitor agent performance and improve the quality of automated decisions. For service providers, the commercial opportunity lies in using AI-driven personalization and data-led engagement to improve loyalty, reduce cost to serve and unlock new value-driven business models.
“This Catalyst solution is valuable because it demonstrates how AI-native, multi-agent architectures can transform customer experience and operations at scale, delivering faster, more accurate, and highly personalized services,” says Luis Díaz Martín, Solutions Architect at Telefónica. “By combining intent-driven automation with composable, interoperable systems, it enables businesses to significantly improve efficiency, reduce costs, and unlock new revenue opportunities. More importantly, it provides a practical blueprint for the industry to move towards autonomous, intelligent platforms that enhance everyday digital experiences for customers and society at large.”
The project draws on a range of TM Forum assets to support interoperability, governance and alignment with industry standards. These include TM Forum Open APIs for structured process orchestration across BSS and OSS systems, TMF921 for AI agent-to-agent interaction, and the Information Framework (SID) and ontology alignment to structure domain knowledge and support semantic consistency.
The Catalyst also aligns with TM Forum’s Autonomous Networks framework, including guidance on security and governance for agentic AI in autonomous networks. The team references IG1463, Security Considerations for Agentic AI in Autonomous Networks, and IG1465, AI for Contact Centre Whitepaper, while building on previous Catalyst work and contributing to broader initiatives around AI-native BSS and AI for business and IT modernization.
Improving experience while scaling trusted automation
By automating routine interactions and improving the accuracy of intent recognition, the Catalyst has the potential to reduce friction for customers while freeing service teams to focus on more complex issues. Faster resolution, higher engagement rates and more personalized interactions can improve customer loyalty, while automation and orchestration can help businesses operate more efficiently.
For the wider industry, the project offers a practical blueprint for moving from monolithic systems towards composable, AI-native architectures. Its focus on interoperability, ontology alignment, governance and multi-agent coordination supports the development of scalable frameworks that can be reused across customer care, sales and operational domains.
The project was named a finalist in the Best Moonshot Catalyst - Composable IT and Ecosystem challenge award at DTW Ignite 2026, recognizing its contribution to the industry’s shift towards composable, AI-enabled business and operating models.
