Amdocs, a provider of software and services to communications and media companies, recently announced significant milestones in its efforts to promote the adoption of generative AI within the telecommunications sector, valued at £1.7 trillion.
This progress stems from a collaboration with NVIDIA. The latest phase of this partnership has achieved meaningful strides towards addressing carrier-scale production imperatives such as latency, accuracy, and token consumption.
Amdocs has closely collaborated with NVIDIA to integrate NVIDIA’s comprehensive AI foundry service into its generative AI platform, amAIz. This integration includes NVIDIA’s AI Foundation Models, NVIDIA NeMo framework and tools, and NVIDIA DGX Cloud, providing enterprises with an end-to-end solution for developing custom generative AI models.
The focus of this collaboration has been on leveraging generative AI to enhance customer experiences and improve operational efficiencies. Amdocs utilised NVIDIA DGX Cloud to test and optimise multiple models, while employing NVIDIA NIM microservices, a component of the NVIDIA AI Enterprise software platform, to rapidly deploy models for seamless, scalable inference using industry-standard APIs.
In joint GenAI projects involving Amdocs’ telco taxonomy and generative AI expertise, the collaboration has achieved the following:
Cost Reduction
Amdocs’ telecom retrieval-augmented generation (RAG) on NVIDIA infrastructure has led to significant reductions in token consumption for deployed use cases. This reduction amounts to as much as 60% in data pre-processing and 40% in inferencing, maintaining the same level of accuracy while reducing the cost per token.
Latency Improvement
The collaboration has successfully reduced query latency by approximately 80%, ensuring near real-time responses for end users. This acceleration enhances user experiences across various domains including commerce, customer care, and operations.
Accuracy Enhancement
The engagement has also resulted in noteworthy improvements in the accuracy of AI-generated responses, with increases of up to 30%. This improvement is crucial for achieving widespread adoption within the telco industry and meeting the demands of generative AI services for direct-to-consumer applications.
“These achievements are not just milestones but are critical in reshaping the telecommunications landscape,” said Anthony Goonetilleke, Group President of Technology and Head of Strategy at Amdocs. “Collaborating with NVIDIA supports our advanced LLM routing capabilities, ensuring that generative AI tasks are matched with the ‘best fit’ LLM, and improving outcomes as a result. Our work speaks to our goal of proving the business value of generative AI, including driving down costs, improving accuracy of results, and delivering tangible benefits to service providers and their customers.”
“By combining Amdocs’ generative AI platform and telecommunications industry expertise with NVIDIA’s full-stack AI foundry service including NVIDIA NIM, we are enabling meaningful scale of generative AI in telcos, accelerating time to value, and driving positive return on investment,” said Chris Penrose, Global Head of Telco Business Development at NVIDIA. “This collaboration is helping to transform how the telecom industry operates, paving the way for a more connected and intelligent future.”
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by commenting below or visiting our LinkedIn page.
