Intelligent Communication Assistant

Description

The Intelligent Communication Assistant (ICA) is a network-resident virtual assistant defined in 3GPP TS 22.870 for Rel-20. Unlike device-centric assistants such as Siri or Google Assistant, the ICA is located within the operator's network and is built upon the IP Multimedia Subsystem (IMS) framework. It interacts with users through multiple modalities—voice, video, text, gestures, or other input methods—and is designed to provide personalized communication services by accessing user data and network data stored or collected within the operator's domain. The ICA represents a significant evolution in how operators can deliver value-added services, moving beyond basic connectivity to intelligent, context-aware communication support.

Architecturally, the ICA leverages the IMS core network as its service delivery platform. IMS provides the session control, media handling, and service triggering capabilities necessary for real-time interactive services. The ICA functions as an application server or a set of application functions within the IMS service layer, capable of initiating, modifying, and terminating sessions on behalf of or in collaboration with the user. Because it resides in the network, the ICA can access a rich set of contextual information, including subscriber profile data from the UDR (Unified Data Repository), network performance data, location information, and service usage patterns. This network-centric placement enables capabilities that device-based assistants cannot easily achieve, such as cross-device continuity, deep network integration, and operator-grade reliability.

The personalization aspect of the ICA is a core design principle. Each subscriber can have one or more ICAs, and each ICA can be customized for the particular user by accessing user data and network data. This customization allows the ICA to understand the user's preferences, communication habits, and typical requirements, enabling it to act proactively—for example, suggesting optimal communication modes, filtering incoming communications based on user context, or automating routine communication tasks. The ICA can also support individual users based on the user's intention and requirement utilizing AI capability, meaning it can interpret high-level user goals and translate them into concrete communication actions.

The ICA is designed to benefit from 6G services when available, specifically referencing 6G AI Services and 6G Computing Services. This forward-looking design ensures that as 6G capabilities are deployed, the ICA can leverage advanced AI models, distributed computing resources, and enhanced network intelligence to improve its functionality. For instance, 6G AI Services could provide the ICA with access to large-scale machine learning models for natural language understanding, while 6G Computing Services could offer edge computing resources to reduce latency for real-time interaction processing.

From a service delivery perspective, the ICA can provide communication services directly—such as setting up calls, managing messages, or orchestrating multi-party sessions—and can also act as an intelligent intermediary between the user and other services. The interaction between the user and the ICA is based on IMS, ensuring that standard IMS procedures for authentication, authorization, session establishment, and media negotiation are applied. This alignment with IMS ensures interoperability with existing network infrastructure and leverages the security and QoS mechanisms already defined for IMS services. The ICA concept is part of a broader trend in 3GPP toward integrating AI/ML capabilities into network services, as seen in other Rel-19 and Rel-20 work items on network automation and intelligence.

Purpose & Motivation

The Intelligent Communication Assistant exists to address the growing demand for personalized, intelligent communication services that go beyond basic voice and data connectivity. Traditional communication services treat all users and sessions uniformly, offering limited personalization and requiring users to manually manage their communication activities. As communication modalities multiply—voice, video, messaging, rich media—and as users expect seamless, context-aware experiences, operators need a mechanism to deliver intelligent assistance that is tightly integrated with the network.

Previous approaches to personalization were largely device-centric, relying on applications installed on the user's device. These solutions have limitations: they cannot easily access network-level data, they are constrained by device capabilities, and they lack the ability to coordinate across multiple devices or integrate deeply with network services. The ICA addresses these limitations by placing the intelligence in the operator's network, where it can access comprehensive data, apply network-level policies, and provide consistent service across devices and access types.

The ICA also aligns with the industry's move toward AI-native networks. By defining the ICA in TS 22.870, 3GPP provides a standardized framework for operators and vendors to develop interoperable intelligent assistant services. This standardization is crucial for ensuring that ICAs can work across different network implementations and can leverage emerging 6G capabilities. The ICA concept also supports the broader goal of network slicing and service differentiation, as ICAs can be tailored to specific verticals or user segments, providing a foundation for new revenue-generating services.