Amdocs Unveils aOS Agentic AI Platform to Speed Telecom GenAI Adoption, Citing Early ROI Gains

Key Points

• Amdocs unveiled aOS, an "agentic operating system" layered on its BSS/OSS with a new Cognitive Core to inject telecom-specific context into agent workflows and drive outcome-focused, measurable ROI rather than just infrastructure investment.
• Early demos and deployments show material gains—Amdocs reports >90% accuracy in automated care, reduced typical call interactions from ~12–14 minutes to ~45 seconds, plus examples like a 135% NPS lift, 40% lower wait times, and a 60% reduction in token usage through tuning.
• Amdocs is leveraging partnerships with AWS, Microsoft, Google, and NVIDIA for cloud, contact-center and network-twin integrations, while testing flexible monetization and adoption models so customers can adopt specific aOS layers based on maturity and regulatory needs.

Amdocs NASDAQ: DOX used a recent investor webinar to walk through how it is positioning itself to help telecommunications providers accelerate adoption of generative AI through a newly launched “agentic operating system,” called aOS, which the company introduced at Mobile World Congress in Barcelona.

Matthew Smith, Amdocs’ head of investor relations, opened the session by stressing that the webinar was intended to be educational rather than focused on financial guidance. Group President of Technology and Head of Strategy Anthony Goonetilleke led the main presentation alongside Pilar (head of global consulting) and Liliana (head of strategic partnerships), highlighting how Amdocs is combining software, services, and partner ecosystems around an “agentic” framework designed for telecom workflows.

aOS positioning and “outcome-driven” focus

Goonetilleke framed aOS as part of Amdocs’ longstanding emphasis on delivering measurable outcomes, rather than “just another technology project.” He described Amdocs as a company that develops “mission-critical” software for telecom operators, and said the complexity and volume of customer journeys, usage events, and billing events handled by telecom systems continue to rise as consumers connect more devices and services.

In describing the broader generative AI market, Goonetilleke said enterprise adoption is “mandatory,” but that customer conversations are shifting from infrastructure investment toward clearer ROI and outcomes. He also emphasized that architecture, governance, organizational readiness, and data privacy have become critical milestones, often outpacing technology as the main constraint.

Inside aOS: cognitive core, telco context, and openness

Amdocs presented aOS as a layered approach built on its existing business support systems (BSS) and operational support systems (OSS), which remain “systems of record.” The company said it has added a new layer, the Cognitive Core, designed to bring telecom-specific context and knowledge into agentic workflows.

Goonetilleke argued that accuracy and latency are central differentiators in telecom use cases. He pointed to the complexity of telecom data and processes—using proration as an example that can be influenced by “over 250 different database entities”—to explain why “context engineering” matters. He also highlighted Amdocs’ long-standing work on a telecommunications ontology via its Amdocs Logical Data Model, saying this helps the system connect business entities such as pricing, policies, customers, and workflows.

He also stressed that aOS is designed to be open, not a “one-stop shop,” describing an “agentic mesh framework” where customers may use different vendors and models. Amdocs, he said, aims to provide reusable components—analogized to “Lego blocks”—that customers can assemble in different ways.

Demos: customer care, billing disputes, and modernization workflows

The webinar included demonstrations of agentic workflows in customer care and IT modernization. In one scenario, Amdocs showed an automated care process triggered by a predicted roaming overage “bill shock” event. The system initiated a chat interaction through a mobile app, retrieved the customer profile, and activated sub-agents to collect usage and financial data from BSS systems. Amdocs described the flow as moving from root-cause identification and bill explanation to a policy-checked dispute resolution and an upsell offer.

Goonetilleke said customers have become more comfortable allowing an agent to respond directly to consumers without a human in the loop, and claimed Amdocs is seeing “90%+ accuracy” in these interactions. He contrasted a typical 12-to-14-minute call center interaction with a response delivered in about 45 seconds in the demo scenario.

Pilar then outlined agentic services, describing them as service domains spanning IT operations, data and AI, cloud, experience design, and quality engineering. She characterized the shift as “services as software,” where expertise traditionally held by experienced staff is codified into orchestrated agent workflows. As an example, she highlighted intelligent application modernization, walking through how humans and agents collaborate—agents doing industrial-scale analysis while humans apply judgment around organizational context and risk appetite. She emphasized that the degree of “human-in-the-loop” control can vary by customer, depending on readiness and regulatory environments.

Early outcomes and cost considerations

Amdocs shared examples of early results it attributed to deployments. In one example using a WhatsApp channel, the company cited a “135% NPS improvement,” increased CSAT, faster payments, and a 40% reduction in wait times. In another example with a tier-one North American provider, Goonetilleke highlighted improvements across call center metrics including average handling time (AHT), first call resolution (FCR), and NPS. He also said Amdocs reduced token usage by 60% over time through tuning, arguing that cost optimization must accompany performance.

Strategic partnerships: AWS, Microsoft, Google, and NVIDIA

Liliana said Amdocs is focusing on four strategic partners—AWS, Microsoft, Google, and NVIDIA—describing the relationships as multidimensional across R&D, engineering, go-to-market, and internal transformation. She provided examples of partner integrations, including embedding Amdocs’ Cognitive Core with contact center solutions and a “customer engagement platform” integrating Microsoft Dynamics 365 with Amdocs BSS capabilities for care, commerce, and CPQ.

She also referenced Amdocs’ work with Vodafone to migrate a complex enterprise data warehouse into Google Cloud’s BigQuery, which she said resulted in faster time to insight, an “AI native data foundation,” and reduced platform costs. Additionally, she pointed to an APAC tier-one deployment of the Microsoft Dynamics 365 and Amdocs combined platform, describing improvements in case management, NPS, and first call resolution for more than 4,000 agents.

On networks, Amdocs discussed a planned demonstration (not fully shown due to time) of work with NVIDIA Omniverse to create a digital twin of the network to support fault detection and autonomous healing.

During Q&A, Goonetilleke described aOS as a re-imagination that incorporates lessons from Amdocs’ earlier amAIz platform, with more “insertion points” for customers across domains such as care and mainframe modernization. He said Amdocs is evaluating monetization models for these capabilities, noting the industry has not reached equilibrium on whether pricing should be token-based or structured differently. He also said customers at different sizes and maturity levels can adopt aOS differently—some taking a broad stack, others integrating at specific layers—depending on where they are in their AI journey.

About Amdocs NASDAQ: DOX

Amdocs NASDAQ: DOX is a global software and services provider specializing in solutions for communications, media and entertainment companies. The company designs, develops and integrates revenue management, customer experience and digital services platforms that enable service providers to launch and monetize new offerings, streamline operations and enhance subscriber engagement. Amdocs' product suite encompasses billing and order management, customer relationship management, digital commerce and network function virtualization, supported by professional services for implementation, integration and managed operations.

Founded in 1982 and structured as a separate public company in 1998, Amdocs has its corporate headquarters in Chesterfield, Missouri, and maintains major development centers in Ra'anana, Israel.

This instant news alert was generated by narrative science technology and financial data from MarketBeat in order to provide readers with the fastest reporting and unbiased coverage. Please send any questions or comments about this story to contact@marketbeat.com.