Amdocs AI Product Capabilities, Professional Services Focus, and Telco Expertise Make it a Strong Contender to Help Telcos on Their Gen AI Journey

Amdocs held its latest analyst summit in London in September 2025, where the company presented its updated telco and enterprise strategy. Among other announcements were the formation of its new Data and GenAI business unit, aiming to help its customers deploy Generative Artificial Intelligence (Gen AI) and Agentic AI at scale. This includes product and service offerings for the following domains:

• Network Automation: The amAIz platform includes several agents for care, sales, network, home, and marketing purposes. It also includes copilots for human operators that can complete commerce, catalog, network operations, and customer engagement functions.
• Vertical Data Platform: Amdocs offers a data platform that can collect, analyze, and store telco data from various data sources, including Operations Support Systems (OSSs), Business Support Systems (BSSs), networks, and others. The vendor claims to be working with all cloud providers (including AWS, Azure, and Google Cloud) and data store/lake vendors (including Databricks and Snowflake) and can utilize these data through any channel.
• Sovereign AI: The vendor has partnered with NVIDIA to utilize its telco platforms for enterprise and sovereign AI use. Amdocs is hoping to help telcos deploy an AI core, which will help them utilize AI in a strategic manner across different areas, including sovereign AI.

Amdocs also recognized that an AI strategy may be disruptive to legacy and current telco operations, and proposed to start this transformation with specific use cases. The company aims to engage with telcos through a consumption-based model, where its customers will pay for outcomes, rather than technology, and will likely deploy its strategy through an overlay on top of existing telco systems.

The AI market is growing rapidly, with training and inference now taking place in massive, centralized data centers in what we can refer to as Phase 1 of AI deployment. However, these data centers are already straining the power grid, which leads to Phase 2, breaking down these data centers into smaller ones and geographically distributing them across a region, where multiple grids are typically present. This helps to solve the problem with the power grid and energy availability. The ultimate transformation step for AI is Phase 3, where inference, and even parts of training, are placed near the end user, in what we can refer to as distributed AI. This is where telcos have a large role to play because they already have real estate (central offices or even cell sites), access to power for these sites, and, in some cases, even the dark fiber needed to connect these sites. However, they do not have the technical expertise, business model understanding, and infrastructure capabilities to take advantage of this trend yet. Indeed, some operators have shifted some of their infrastructure to the cloud, but have not embraced truly cloud-native processes that are necessary for AI.

Telco transformation for AI can take place either top-down or bottom-up:

• In a bottom-up transformation, telcos deploy Graphics Processing Units (GPUs) in their real estate and start gradually adapting the upper layers of their stack to monetize and orchestrate this infrastructure. This is a Capital Expenditure (CAPEX)-heavy approach, and would have to follow clear market demand, especially as telcos cannot invest proactively in a market they have little experience in. Some telcos are indeed investing in AI data centers, or partnering with neocloud providers, but most are not fully transforming their operations for AI.
• In the top-down approach, telcos start transforming their control infrastructure, namely OSS/BSS. This represents a more moderate effort and allows telcos to adapt the upper layers of their stack—which is software—to the growing AI market. Deploying GPUs in their network would come at a much later stage, but some operators—like SoftBank—have already started doing so.

Most telcos are following the top-down approach, but many do not have the resources or expertise to manage this transformation.

The OSS/BSS ecosystem is uniquely positioned to help telcos transition into the AI era. Vendors such as Amdocs, Netcracker, Ericsson, Nokia, and Oracle have longstanding commercial relationships with operators and deep expertise in network data, charging, policy, Customer Relationship Management (CRM), and service catalogs. Many of these layers are now being enhanced with AI, including Amdocs’ amAIz platform for customer care, Netcracker’s GenAI TelcoOps, Nokia’s AVA AI, and Ericsson’s Cognitive Network Operations.

A key differentiator for Amdocs lies in its professional and managed services capabilities. Each operator’s environment is unique, shaped by decades of legacy systems, multi-vendor dependencies, and diverse technology stacks. As the AI market evolves rapidly, deploying AI across networks will require complex integration efforts—spanning Large Language Models (LLMs), GPU-as-a-Service (GPUaaS) infrastructure, sovereign AI clouds, Neural Processing Units (NPUs), and heterogeneous compute architectures (from NVIDIA and Intel to Arm-based and domain-specific accelerators).

The next wave of telco transformation will, therefore, be integration-intensive, multi-domain, and AI-driven. This is where Amdocs and other OSS/BSS vendors can lead—by positioning themselves not merely as a BSS/OSS vendor, but as the strategic AI systems integrator for telcos and by turning integration complexity into a managed capability.