Translating customer intent into assured service outcomes

Enterprises increasingly expect connectivity services that can adapt to changing business needs quickly and intuitively. Yet ordering and assuring advanced services still often requires customers to translate business outcomes into technical specifications, while CSP teams manually validate feasibility, resolve issues and redesign services as network conditions change. For CSPs, this creates long lead times, rework, inconsistent outcomes and high cost to serve at a time when they are under pressure to monetize capabilities such as latency, resilience, and security.

The Catalyst project, Agentic I2O – Intent to outcome, addresses this challenge with an intent-driven, agent-assisted architecture that lets customers express the outcome they need, rather than the technical requirements behind it. Customer intent is decomposed across business, service and resource layers, validated in real time against commercial rules and live network conditions, and then translated into feasible, assured services using open TM Forum standards.

The challenge of turning intent into action

Modern networks span legacy and cloud-native infrastructure across access, transport, core and edge. Although CSPs already have rich telemetry and programmable capabilities, many fulfilment and assurance processes remain fragmented, policy-driven, and manually orchestrated. Customers are often forced to explain requirements repeatedly across channels, while CSPs must interpret those needs, validate what is possible, and intervene when delivered services no longer match the intended outcome.

The commercial case is equally clear. Traditional product-centric models make it difficult to price and deliver connectivity based on assured outcomes. By shifting the focus from static service definitions to customer intent, CSPs can reduce avoidable rework, lower repeat contacts, and create differentiated offers based on assurance level, resilience and adaptability.

A new model for service delivery

The Catalyst introduces a model in which agentic AI interprets business intent, evaluates feasibility, and selects the best realization path. If constraints arise, the system can propose outcome-equivalent alternatives instead of pushing the customer back to the beginning of the ordering process. Once activated, closed-loop assurance monitors the intended outcome, not just static service-level agreements, and can autonomously adapt the service, for example by upgrading to higher-assurance connectivity when conditions change.

This approach supports zero-touch fulfilment and assurance by combining real-time feasibility checks with autonomous service adaptation. The project aims to reduce service delivery cycles from weeks or months to minutes or hours, shorten revenue realization timeframes, and prevent outcome degradation before customers need to raise support requests.

From natural language to assured outcomes

The project uses a broad set of TM Forum assets to demonstrate a standards-based path from intent to outcome. These include IG1253 Intent in Autonomous Networks, TR290 Intent Common Model, TR291 Intent Extension Models, TR292 TM Forum Intent Ontology, TR293 Connector Model, TMF921 Intent Management API, TMF921A Intent Management API Profile, TMF622 Product Ordering API, TMF641 Service Order Management API, TMF642 Alarm Management API and TMF645 Service Qualification API.

Together, these assets enable a reusable architectural framework for translating high-level intent into complex network services across domains. They also help demonstrate how agentic AI can be applied selectively to reduce customer and operational effort, supporting the industry’s progress toward Level 4 autonomous network capabilities.

Measuring success

The Catalyst’s impact is focused on customer effort, speed, cost, and trust. By validating feasibility up front and maintaining outcomes autonomously, the solution targets common sources of friction such as repeated interactions, re-explaining requirements and misalignment between what customers ask for and what is delivered. The team also points to customer effort research showing that customers who report low effort have a 94% likelihood to repurchase, an 88% likelihood to increase spend and only a 1% likelihood to spread negative word of mouth.

Why it matters

For businesses, this can mean faster time to value, lower cost to serve, and new outcome-based monetization opportunities. For the wider industry, it provides a practical example of standards-based autonomy using intent management and multi-agent systems. For society, the project points to more reliable digital services, greater transparency between customer intent and service behavior, and reduced waste through dynamic cross-domain optimization rather than static over-provisioning.

Richard Keele, Product Owner at Telstra, says the Catalyst directly supports Telstra’s Connected Future 30 strategy by making it simpler for customers to access increasingly sophisticated connectivity while accelerating progress toward autonomous, programmable networks. He explains that intent-based designs allow Telstra to treat the network as a product with clear commercial and customer value, automating how customer needs are translated into assured outcomes across its Network as a Product and Customer Engagement layers.

DTW Ignite 2026

Agentic I2O – Intent to outcome was showcased at DTW Ignite 2026 as part of the Catalyst program, demonstrating how intent-driven, agentic service delivery can help CSPs turn business requirements into assured outcomes with far less friction and far greater automation. The project was named the Open Innovation Catalyst winner in the Customer experience and reinvention category.