'Everything is becoming AI intelligent and AI-enabled': Oracle database head says agents are the future, but warns “there's no magic bullet”

Agents are becoming an increasingly common presence in many businesses, helping human workers become more efficient and productive.

As one of the world’s most prominent data firms, Oracle is looking to help lead the way when it comes to greater AI agent integration in the workplace, and at its recent Oracle AI World Tour London event, revealed a host of new releases and tools in that vein.

But what will this AI agent-powered future actually look like? I sat down with Juan Loaiza, EVP, Oracle Database at the event, to find out more.

The AI world

“In the AI world, you’re on a very rapid treadmill,” Loaiza tells me, “the technology is advancing so quickly, and the adoption is so fast, it's like nothing we've ever seen before.”

“AI is moving extraordinarily rapidly, and so we have to keep up with that speed,” he notes, “there's a lot of parts to this thing! Everything is agentic now...so we're putting a lot of work into agents, in many different ways.”

“Clearly agents are the next big thing in AI - so we want to be able to build very data-centric agents very easily,”

Oracle revealed its AI database launch in October 2025 in a move Loaiza says re-architected the company’s whole data platform around AI, and the ability to easily build AI-centric agents is clearly a major priority for the company going forward.

Loaiza notes how “every piece of the Oracle data stack, and every component, is becoming AI-enabled”, with over 100 features in its database empowered that way right now.

“Everything is becoming AI intelligent and AI-enabled.”

But a big part of greater acceptance of AI will be trust, Loaiza notes, describing it as, “the fundamental building block everything else has to be built on.”

Especially as workers increasingly look to harness AI to improve their productivity and efficiency, Oracle is looking to ensure private data, whether personal or enterprise, is kept protected from AI tools, right down to the data level, which cannot be bypassed. Customers can now define rules on what database access is permitted, so when any data is changed, the mechanism will kick in, making sure that when the AI modifies data, there is no chance of error or corruption - hopefully.

“There's no magic bullet,” Loaiza admits, “so what we're building is different ways to enhance trust for different kinds of workloads and different kinds of users.”

Loaiza outlines three levels of trust, going from that needed from an experienced workplace user such as a software developer, to more general worker, such as an HR employee, down to a member of the public.

“What we need to understand is, the goals are different at each of these levels,” he says, “we're starting to learn that AI is not one uniform thing - you have to separate the different use cases, they have different requirements, different levels of trust that are required - and that's part of the whole learning process of the new world that we're entering.”

Loaiza admits trust is “the big challenge going forward” for the level of trust given to AI, noting that even natural language questioning can be ambiguous or ill-defined, so AI might get the wrong response due to misunderstanding your question

There's no doubt AI can offer huge benefits to workers everywhere - Loaiza says he sees the main benefits of the technology spreading across productivity, innovation and getting insights and answers from data - saying the technology could unlock, “almost a different world” unlocking possibilities for all end users.

“Since computers were invented, humans have had to learn the language of computers in order to interact,” he adds, “now computers can interact with the language of humans - that takes it from 0.01% of the population being able to interact effectively with computers and data, to 100% of the population!”