Anthony Goonetilleke, Group President, Technology, and Head of Strategy, Amdocs
Generative AI is evolving more rapidly than any technology in my lifetime. Never before have we seen so many industries and experiences impacted almost simultaneously. The relentless march of innovation in this field is breathtaking, and the potential applications seem limitless. As I write this article, I am acutely aware that in the week between authoring and publication, there will be new developments, new models introduced, and pricing models will continue to evolve. That's the pace at which generative AI is advancing, and it's a testament to the incredible potential it holds.
The Evolution of Generative AI
At its core, generative AI is a super promising technology, with wide-reaching potential for adding value to nearly every vertical, from academics to healthcare, logistics to manufacturing, and beyond. We talk about advancements in connectivity driving more equitable access, and generative AI only stands to 10x that evolution by making knowledge more widely accessible.
At Amdocs, we believe in the power of generative AI to introduce transformative change in telecom – from simplifying the operations of networks to streamlining interactions with customers, and much more. The technology has huge potential to create more agile organizations as it is embedded in daily processes ranging from back-office options to end-user experiences.
Our generative AI framework, amAIz, is a critical framework for our industry, as it simplifies the requirements for our customers to adopt generative AI in their businesses. By leveraging the best of the industry’s evolving foundational LLMs, and overlaying a robust vertical taxonomy and pre-configured ‘use case kits,’ our customers simply need to choose what generative AI capability they’d like, and we take care of all of the important underlying technology integration, training and governance.
The Promise of Generative AI
One of the remarkable aspects of the evolution of generative AI is the sheer scale of the models being created. There are countless numbers of parameters being added to these models, and we are starting to see compelling instances of multimodal generative AI emerge. This is a significant leap forward, as it allows AI systems to understand and generate content across multiple sensory modalities, such as text, images, and even sound. Imagine a future where AI can not only read text but also comprehend images and interpret spoken language with equal proficiency – that's the kind of potential we're dealing with.
However, as with any technology change, governance, security, observability, and scalability are critical components. Service providers need to ensure that the benefits of generative AI are harnessed safely and effectively. At Amdocs, we're investing heavily in ensuring that service providers can rapidly and safely adopt generative AI in nearly every aspect of their business. We understand the importance of striking the right balance between innovation and responsibility.
What excites me the most about the current state of generative AI is that this is just the beginning. We don't know what we don't know yet, and that's an exciting place to be. Companies like Meta, OpenAI, NVIDIA, Microsoft, and others are making tremendous progress on defining what comes next – and we are thrilled to be a part of this journey. The collaborative spirit in the AI community is propelling us forward at an unprecedented pace.
Scaling Up and Looking Ahead
The future is surely an exciting place to be. A recent report from Gartner indicated that in little more than a year’s time, 10% of all data produced will come from generative AI. Another study from McKinsey estimated that software engineers will experience a productivity improvement of more than 20%. Think about what that means to your business; the tremendous potential to make massive improvements in the way we operate, but also consider the significant risk if we don’t do it correctly.
I traveled recently to visit with some of the biggest names in generative AI. The projects and capabilities they were working on floored me. Remarkable technology, even more remarkable outcomes. Seeing multimodal LLMs in the flesh is a technologist’s dream come true, and it’s just around the corner. They’re also conducting more aggressive accuracy correction techniques, like mid-training, where the model efficiency can be improved vastly, particularly as retrieval augmented generation techniques are further honed.
And of course, we can’t talk about generative AI without mentioning hallucinations (remember my comment about doing it correctly?). Leveraging LLMs to realize all of this promise hinges on the optimization of models to substantially avoid hallucinations. By introducing techniques like Chain-of-Thought (CoT) and Chain-of-Verification (CoVe) prompting, and grounding with RAG (and eventually visual grounding for mult-modal LLMs), I could begin to see how this will all play out as the technology continues to mature.
In the end, generative AI is a transformative force, and its impact is much larger than technology. Organizations, and industries, need to consider everything from the talent they are developing, to the tools they’re using, to the targets they’re setting, and put an intentional focus on how generative AI will best serve their people, their organization, and their customers. The future is nearly here, and you’re either riding the wave, or you’ve already missed your shot.
