Speaking on the evolution of enterprise AI, Sholay points to a decisive move towards “systems of outcome”, where AI agents actively execute tasks rather than simply support them.

“You’ll be able to say, ‘go and get me the best prices, past purchase history, and help me negotiate’ and the agent may even do the negotiation on your behalf,” he explains. The result is a fundamental shift from dashboards and manual workflows to dynamic, outcome-driven interactions.

For Oracle, the value of this shift is measurable. Sholay outlines a framework built around three core pillars: efficiency, effectiveness and experience. “Efficiency is about saving time. Effectiveness is about doing the job better. And experience—if it’s hard to use, it simply won’t deliver value,” he says. These metrics underpin Oracle’s AI value model, enabling organisations to compare performance before and after AI deployment.

Crucially, this measurement is grounded in real operational data. Within Oracle’s applications, business and task metrics are continuously tracked, allowing companies to quantify return on investment with precision. Sholay offers a practical example: reducing time-to-hire from 68 days to 51 through embedded AI agents. “You now have a definitive metric—you can clearly see the improvement,” he notes.

While cost savings remain the most immediate benefit, particularly through productivity gains, Sholay believes the real opportunity lies elsewhere. “The market is moving towards entirely new AI-driven products, services and experiences,” he says. This is where the distinction between “horizontal” and “vertical” AI becomes critical.

Horizontal AI, such as copilots layered onto existing processes, delivers incremental gains. Vertical AI, however, reimagines entire business functions. Sholay highlights a complex fleet management use case involving electric vehicles. Traditional route planning, managed overnight by teams, is replaced by real-time AI optimisation that factors in traffic, battery levels, charging availability and cost. “You can replan and optimise routes every five seconds,” he explains. “The business impact is orders of magnitude greater.”

Such applications demonstrate AI’s ability to transform not just efficiency, but decision-making itself, compressing hours into minutes while improving customer outcomes and avoiding costly penalties.

Despite this progress, barriers remain. Data accessibility is a primary challenge, particularly when critical datasets remain on-premise and not AI-ready. Equally, many organisations are still focused on low-impact use cases. “Chatbots and assistants are valuable, but companies need to solve real business problems,” Sholay argues.

He is also candid about organisational mindset. “AI has been treated as a tool owned by IT. That needs to change. It should be driven by the line of business: the people who understand the problems and the budgets,” he says. This shift, alongside clearer outcome-focused strategies, marks a growing maturity in enterprise AI adoption.

Looking ahead, Sholay expects many of today’s AI capabilities to become embedded within core platforms. “What feels like magic today will soon be standard functionality,” he observes. This belief underpins Oracle’s approach of integrating AI into its applications at no additional cost, positioning it as a foundational capability rather than a premium add-on.

Ultimately, the message from Sholay is that the era of AI pilots is ending. What matters now is scale, impact and measurable value. As Sholay puts it, “Start with a business problem, prove it works, and then scale. That’s where the real transformation happens.”

RELATED STORIES

Oracle targets ‘autonomous’ enterprise operations with new agentic applications

Oracle adds new AI services to $5bn plans for UK government and defence sectors