Exclusive: How Oracle integrates AI agents to boost enterprise productivity

Techday NZ

21-07-2025

In an era defined by rapid advances in artificial intelligence, the competitive arena of enterprise software is being reshaped by the emergence of sophisticated AI agents.
Oracle, a long-standing giant in business technology, is positioning its AI-powered agents at the heart of its application suite, promising deep productivity gains, richer insights, and a fundamentally new way of working for global enterprises.
"The key about agents, from our point of view, comes down to four things: language, context, the ability to use tools, and reasoning," said Miranda Nash, Group Vice President responsible for Oracle's AI strategy, during a recent interview.
According to Nash, these AI agents contrast sharply with traditional automation tools.
"What differentiates agent-based automation from past phases is reasoning, so the reasoning capability of the models means that some of the decision making can be left to the model. This allows us to go much, much faster and have better versatility in the face of real world, unexpected conditions."
The transition to this new phase is underpinned by Oracle's deep integration of AI within its Fusion Applications suite.
Instead of siloed bots or bolt-on assistants, AI agents are embedded directly into workflows across finance, HR, supply chain, and more.
Nash underscores that customers receive these enhancements automatically through regular application updates, meaning that "they have all this capability at their fingertips… already there in the products they own today, and they need to enable it and take advantage of it."
The measurable impact is already being felt by enterprises.
A notable example cited by Nash involves a large hospitality customer that saw goal completion rates within HR processes double after deploying embedded AI agents.
"We've seen basically a two times improvement in terms of the completion rate. Without AI versus with AI, folks are 100% more likely to complete the task, which means you're getting better employee engagement and more value out of your existing systems."
This translates not only to improved productivity but to real gains in employee engagement and operational value.
The technical sophistication of Oracle's agents lies in their multi-dimensional approach.
Agents can ingest large volumes of private enterprise data for contextual answers, invoke relevant APIs, and-critically-reason with that data to facilitate complex decisions. Nash compares the experience to having a "super powered Chief of Staff" operating continuously.
"Able to gather information that you need from lots of sources," Nash says, these agents unlock latent productivity and creativity among employees, who discover new ways to apply AI beyond their initial expectations.
With the launch of Oracle AI Agent Studio, much of this power is now placed in the hands of customers and partners, enabling them to create, extend, and deploy custom AI agents tailored to their unique processes.
The platform leverages a library of pre-built agent templates, orchestration tools for managing agent teamwork, and the ability to connect with both Oracle and third-party systems.
Nash highlights that these tools arrive with "built-in validation and testing" to help maintain trust and accuracy in AI-driven workflows.
"They are built in right where the work is happening, not bolted on in some separate system-that gives our customers a better user experience, faster adoption, and, of course, guarantees the type of security and data protection that you would expect from Oracle."
Security and trust, especially around data privacy, are embedded in the DNA of Oracle's approach.
"Anytime we are invoking an agent, it is using the security profile of a human user," Nash explains. This means strict adherence to existing role-based access controls.
"An agent that is working on my behalf will only have access to the data that I'm allowed to see." Oracle also maintains a chief privacy officer, with AI subject to the same contract and data control requirements as any other technology in its suite.
The governance of AI, especially as agents grow more autonomous, remains a top consideration.
Nash describes how Oracle provides granular controls, allowing customers to determine when and where human oversight is required in automated processes.
"In our AI agent studio, there's basically a radio button to have human oversight or 'human in the loop' on any process. We try to make it very easy and give the customer control, because all customers have a different trust profile in terms of how quickly they come to trust the AI," she notes.
The envisioned lifecycle moves from manual oversight to growing degrees of autonomy as trust builds.
For users across industries-including finance, hospitality, manufacturing, energy, and even regulated sectors-adoption has, in some cases, moved faster than anticipated.
"Regulated industries, which you might not expect, are actually some of the more aggressive adopters of AI, because they already have in place a governance infrastructure, so they didn't have to start from scratch," Nash observes.
Despite the intense industry buzz surrounding AI, Nash points to a surprising challenge: "Many of our customers are still not aware that they have all this capability at their fingertips."
Oracle has responded by deploying large teams across the business to assist customers in surfacing and activating these new AI-driven features.
In the future, Nash anticipates a significant evolution in the sophistication of AI agents within the enterprise.
"We will see users gain more trust, and with that increased trust, we will see agents taking on more autonomy, meaning there is no human in the loop for low-risk decisions."