Artificial intelligence is entering a new era of enterprise dominance, with agentic AI emerging as the catalyst behind sharper efficiencies, cost savings and elevated customer experiences.

Agentic AI is becoming a driving force for businesses aiming to reduce fraud, streamline operations and achieve measurable returns in competitive markets. By fine-tuning large language models to meet specific objectives, organizations can cut costs without sacrificing accuracy, while also democratizing access to AI tools — reshaping company culture and empowering teams to innovate more efficiently, according to Andy Markus (pictured, left), chief data and AI officer at AT&T Inc.

AT&T’s Andy Markus talks agentic AI with theCUBE.

“We’re building on a very rich history of traditional AI with generative AI and agentic AI. We have a very disciplined program,” Markus said. “We take use cases across the company, we bring those to a central pool, we evaluate those with a very close peer of mine on the business side and we do a business case. We returned two times ROI free cash flow impacting ROI on every dollar that we spent with generative AI last year.”

Markus was joined by Sri Satish Ambati (center), co-founder and chief executive officer of H2O.ai Inc., and Andrew McMullan (right), chief data and analytics officer of Commonwealth Bank of Australia, as they spoke with theCUBE’s Savannah Peterson at the Nvidia GTC event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how agentic AI and fine-tuned large language models are driving operational efficiency, fraud prevention, cost savings and cultural transformation within enterprises while delivering measurable ROI and empowering non-technical teams to leverage AI for business innovation. (* Disclosure below.)

Democratizing agentic AI across the enterprise

Enterprise leaders are prioritizing the integration of agentic AI throughout their organizations, making advanced technologies accessible to non-technical teams. By training thousands of employees and encouraging widespread usage, they’re fostering a culture where innovation is not confined to data scientists alone, according to McMullan.

“Over the last 12 months we actually trained 900 of our analysts to use H2O,” he said. “What that means is … we talked about fraud and scams but you think about every decision we make in our customers and we make millions of decisions every day. In every one of those we’re making those decisions 100% better using H2O.ai than the previous models.”

This democratization is also driving a shift in business strategy, enabling faster product releases and better customer outcomes. Companies are finding that when teams at all levels are empowered with AI tools, the pace of innovation accelerates, Markus added.

“Everything that we do is based on an ROI basis,” he said. “We look at the value that we’re creating and what we’re spending to do that and really prioritizing what makes the difference for AT&T and our customers. Everything from the experience on the digital side, from the experience in the call center, to just how we do things more efficiently to run the network better, all those things are now driven with agentic solutions and generative AI solutions to make that process better.”

Agentic AI drives real-world impact and ROI

The measurable benefits of agentic AI extend far beyond internal efficiencies. Enterprises are leveraging these solutions to deliver meaningful societal impact, from fraud prevention to environmental sustainability, McMullan explained.

“Over the last 12 months, we’ve reduced fraud losses by over 30% for our customers. We’ve reduced scam losses by over 70%,” he said. “That’s using the great models that Sri and the team come up with just to help our customers have a better experience.”

At the same time, careful prioritization ensures that AI investments yield tangible returns. Formal business cases, combined with continuous innovation, are allowing companies to stay ahead of competitors and bad actors alike, according to Ambati.

“Customers invested in us when no one else believed in AI back in the day,” he said. “Intelligence is ubiquitous, agency becomes very important. If something’s calling you, follow it. Then maybe the world will open up in there and AI will be that true transformation.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the Nvidia GTC event:

(* Disclosure: H2O.ai Inc. sponsored this segment of theCUBE. Neither H2O.ai nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE