Altair Newsroom

Featured Articles

Future Says S4E8: Pushing the Envelope with Generative AI

Viji Krishnamurthy, senior director for generative artificial intelligence (AI) solutions at Oracle, is undoubtedly well suited to share insight on the impact of the transformative technologies impacting today’s commercial world. Since completing her doctorate in operations research at Northwestern University, Krishnamurthy has demonstrated an impeccable ability to catch the crest of some of the most powerful waves that have swept through industry in recent years. Initially, this ability guided her to a data science role at Philips before moving to Oracle for the first of two roles focused on the Internet of Things (IoT). The next step, and technology shift, was to lead Oracle’s AI services team, a position that has evolved into one that makes her responsible for exploring how generative AI can reimagine business processes. 

Given Future Says’s ongoing mission to share expert perspectives on data science and AI, Altair was delighted to welcome Krishnamurthy as the guest for the eighth and final episode of Future Says season four. In this episode, she sheds light on how generative AI can add value to organizations today. She also predicts where the technology is heading and the challenges that developers will need to overcome if they want to push the envelope further still. 


From Prediction to Creation

The business community at large shares Krishnamurthy’s passion for generative AI. According to a recent report by Bloomberg Intelligence, the market for generative AI is set to grow from “just” $40 billion in 2022, to an astonishing $1.3 trillion within a decade. That’s a compound annual growth rate (CAGR) of 42%.

As Krishnamurthy explains, this isn’t just hot air – there’s real substance behind the enthusiasm. For enterprises, generative AI represents a profound leap forward. To date, businesses have generally employed AI to explore historical data for patterns and to use those patterns to predict likely future outcomes. As such, AI fosters better decision-making. Generative AI is different because, as the name implies, it’s capable of creating new content, even from natural language prompts and instructions. 

For Krishnamurthy, generative AI opens the door to a myriad of ways organizations and individuals can supercharge productivity. These include the ability to create first drafts of documents, extract relevant information from vast data resources, and present and summarize information rapidly in a way that makes it accessible to any possible audience. And these are just a few examples.


Harnessing Knowledge

One of generative AI’s biggest advantages is that it has the potential to put an organization’s entire knowledge base at each employee’s fingertips. A typical manufacturer, for example, develops a huge well of information that covers things like how equipment works, how it’s maintained, how products are built, and more. Currently, that knowledge is likely distributed across an array of lengthy documents and/or the company intranet. Generative AI can make it instantly accessible via natural language prompts and responses. That means it’s possible to empower all engineers, for example, with best-in-class competencies. As Krishnamurthy explains, the question we should be asking is: “Irrespective of whether someone has only five years of experience or 25 years of experience, can we level set their knowledge?” 

But that’s just one potential avenue to explore. Krishnamurthy believes that company executives need to be asking even more questions:

  • How do employees currently spend their time?
  • How can generative AI better enable and empower them?
  • How can generative AI accelerate routine tasks?

And Krishnamurthy also insists that generative AI is ready to take on these processes now. At the same time, there’s a real need to look further ahead. “Organizations should think about generative AI journey as a long game,” she says. Planning should also embrace the fact that generative AI will undoubtedly become more accurate, more reliable, and more consistent. However, that shouldn’t be used as an excuse for delaying investment. 


Looking for the Edges

For anyone who wants to keep pace with such a fast-evolving technology, Krishnamurthy’s advice is clear. “You have to constantly learn,” she says. “You have to always look at what is happening now and also think, what are the edges of the envelope, where is it going to break?” What does that generative AI “envelope” look like? To her, it’s comprised of the data, the compute, the algorithmic framework, and the use cases.

In terms of pushing the boundaries, Krishnamurthy highlights two particular challenges. The first is that the current iteration of generative AI models are “instinctive.” This means the models are good at predicting the next step in a process (such as the next word in a sentence), but don’t replicate the more nuanced decision-making that humans excel at. One of the ways this hurdle might be overcome is by getting large language models (LLMs) to work with one another. But that’s another key weakness in today’s generative AI. “We still don’t have a good way of making one LLM work with another LLM to make a decision.” 

Over the next year or so, however, Krishnamurthy thinks we’re likely to see significant progress in these areas, which will open even more possibilities to create positive change. In her own words, “the future is bright.” 

Click here to listen to the full episode with Oracle's Viji Krishnamurthy. To learn more about the Future Says series and browse previous episodes, visit

Future Says Season 4 is proudly sponsored by Oracle. Oracle offers integrated suites of applications plus secure, autonomous infrastructure in the Oracle Cloud. For more information about Oracle (NYSE: ORCL), please visit


S4 Previous Recap Articles