The Future of AI: What Comes Next, and What It Means for Oracle Professionals
From AI assistants to autonomous agents, the next wave of innovation could fundamentally change how we build applications, manage data, and run enterprise systems.
Artificial intelligence has moved a long way in a relatively short time. We’ve gone from experimenting with chatbots and asking AI to write snippets of code to building applications, analysing complex datasets, and automating tasks that previously required considerable human effort.
But I think we’re only at the beginning.
The next stage of AI will be less about asking questions and getting answers, and more about giving AI systems the ability to carry out meaningful work. For Oracle professionals, developers, DBAs, architects, and enterprise technology teams, that presents some
exciting opportunities, along with a few challenges we need to start thinking about now.
So, what might the next few years bring?
1. AI agents: From answering questions to getting things done
One of the biggest developments to watch is the rise of AI agents.
Most of us are familiar with using AI to explain a technical problem, generate SQL, write documentation, or help troubleshoot an error. AI agents take this a step further. They can work towards a defined goal, use tools, retrieve information, execute multiple steps, and coordinate tasks across different systems.
Imagine asking an AI agent to investigate a database performance issue. It could review monitoring data, examine execution plans, compare recent changes, consult documentation, and prepare a remediation plan. With the right permissions and safeguards, it might even carry out approved corrective actions.
This is where things get interesting for enterprise technology.
Oracle is already moving in this direction. Its Oracle AI Database innovations include agent-building capabilities, integration with enterprise data, and mechanisms for agents to interact with database services. Oracle’s announcements around Private Agent Factory and agent-to-agent communication demonstrate how the technology is evolving beyond standalone chat interfaces.
For more information, see Oracle’s introduction to AI Database Private Agent Factory.
I expect we’ll see more organisations experimenting with agents for database operations, service management, application support, reporting, and business workflows.
The important distinction is between an impressive demonstration and a reliable production system. Agents will need carefully controlled permissions, clear boundaries, audit trails, and human approval for potentially disruptive actions.
2. Oracle Database becomes an even more important part of the AI story
There’s a common misconception that AI is primarily about choosing the best language model. In an enterprise environment, the data behind the answers is often just as important.
A model might be very capable, but it won’t automatically understand an organisation’s customer records, business rules, application metadata, or internal documentation.
That’s why I think the integration of AI with enterprise data will be one of the most significant developments over the coming years.
Oracle AI Database brings together relational data and capabilities such as vector search, allowing developers to build applications that combine conventional database queries with semantic searches and AI-generated responses.
This opens up possibilities such as:
Searching enterprise documentation using natural language.
Finding relevant information across large collections of business records.
Building assistants that answer questions using approved organisational information.
Combining structured database results with unstructured documents.
Developing AI agents that retrieve business data and initiate controlled workflows.
For example, imagine a support assistant that can search technical documentation, identify relevant historical incidents, retrieve information about an affected system, and suggest a solution based on the available evidence.
Rather than replacing the database, AI makes the information held within it more accessible and useful.
For organisations already invested in Oracle, this could provide a practical route to introducing AI without abandoning established data platforms.
3. Application development is about to change again
Another area I expect to see substantial change is software development.
AI coding assistants are already useful for generating SQL, PL/SQL, JavaScript, and other code. But the next step is to move from generating individual pieces of code towards helping design, build, test, and maintain entire applications.
Oracle APEX is particularly interesting in this space because it already provides a low-code environment for building data-driven enterprise applications.
Oracle’s AI-powered APEX development capabilities are expanding the ways developers can use natural-language instructions to create applications and work with existing code. Its APEX 26.1 direction also includes APEXlang, a structured way to represent application definitions that can be reviewed and governed.
You can explore the current capabilities at Oracle APEX and AI.
I don’t think this means developers will suddenly become unnecessary. Quite the opposite: I think the role of a good developer will evolve.
Writing routine code may become less time-consuming, but understanding business requirements, designing sound architectures, securing data, integrating systems, and validating results will remain essential.
The developers who learn how to combine AI-generated solutions with sound engineering practices will be in a strong position.
4. Enterprise integration will become even more important
As AI agents become more capable, they will need to interact with an increasingly diverse collection of enterprise systems.
An AI assistant that can only answer questions is useful. An assistant that can securely retrieve information from several systems, coordinate a workflow, and return a verified result is potentially much more valuable.
This creates opportunities around APIs, integration platforms, identity management, data access, and orchestration.
For Oracle professionals, that means keeping an eye on how AI connects with Oracle Integration, Oracle Cloud Infrastructure, databases, enterprise applications, and third-party services.
Standards and protocols that allow AI applications to discover and use tools will also become increasingly relevant. Oracle’s work on agent-to-agent communication is one example of this direction.
The challenge will be making these connections reliable and secure. Giving an AI agent access to business systems introduces new questions about permissions, authentication, data leakage, monitoring, and accountability.
The organisations that solve those problems well will be better positioned to move beyond experimentation.
5. AI governance and security will move up the priority list
As AI systems gain the ability to take action, governance becomes much more than a policy document.
Organisations will need to understand what their AI systems can access, which decisions they can make, what actions they have performed, and how those actions can be reviewed or reversed.
Consider an agent that can query a database. Should it be able to see every table? Can it access personal information? Is it allowed to execute updates? Who approves changes to production data? How do we investigate an unexpected result?
These are familiar questions for anyone working in enterprise security and operations, but AI introduces additional complexity.
The 2026 Stanford AI Index highlights the continuing gap between advances in AI capability and progress in responsible AI practices.
For technology teams, I think this creates an important opportunity. People who understand security, data governance, identity, auditing, and operational risk will have a significant role to play in making enterprise AI trustworthy.
Getting AI into production is one challenge. Making it safe enough to operate there is another.
6. Scientific discovery and robotics could deliver some surprising breakthroughs
Beyond enterprise IT, AI could accelerate innovation in fields such as medicine, materials science, energy, and manufacturing.
AI models are increasingly being used to identify promising research directions, analyse experimental results, and help scientists design new experiments. Combined with automated laboratories, these capabilities could shorten parts of the research process.
Robotics is another area worth watching. As AI becomes better at interpreting its surroundings and responding to changing conditions, robots should become more useful in warehouses, factories, agriculture, and other controlled environments.
I expect practical industrial applications to develop faster than general-purpose household robots.
These advances are harder to predict than software developments because physical systems must work reliably in the real world. Nevertheless, the combination of AI, automation, and scientific research could have consequences far beyond the technology industry.
7. What should Oracle professionals do now?
With so many developments happening at once, it can be tempting to chase every new tool or model. I think a more practical approach is to focus on capabilities that will remain valuable as the technology changes.
Here are five areas I’d be paying attention to:
1. Learn how AI works with enterprise data. Understand retrieval-augmented generation (RAG), vector search, embeddings, and the challenges of grounding AI responses in reliable information.
2. Experiment with AI agents. Learn how agents use tools, call APIs, coordinate tasks, and operate within defined permissions. Start with a small, low-risk use case rather than giving an agent unrestricted access to production systems.
3. Explore AI-assisted development. If you work with Oracle APEX, SQL, or PL/SQL, experiment with AI-assisted development and test whether it improves your productivity without compromising code quality.
4. Strengthen your security and governance knowledge. Identity, access controls, data protection, auditability, and human oversight will be essential as AI systems gain more autonomy.
5. Keep your underlying technical skills sharp. Understanding databases, application architecture, integration, troubleshooting, and business processes will help you recognise when AI-generated solutions are correct—and when they aren’t.
The goal isn’t to become an expert in every AI product. It’s to understand how these technologies can solve real problems in the environments we work in.
Looking ahead
My prediction is that the next few years will see AI move from being an optional productivity tool towards becoming an integral part of many enterprise systems.
We’ll see more capable agents, tighter integration with business data, faster application development, and increasing automation of routine operational tasks.
However, progress won’t be uniform. Reliability, cost, security, regulation, and organisational readiness will determine how quickly individual technologies become mainstream.
For the Oracle community, I see a particularly interesting opportunity. We already work with the databases, applications, integration platforms, and enterprise processes that AI needs to become genuinely useful.
The challenge—and the opportunity—is learning how to connect those foundations to new AI capabilities while maintaining the reliability and trust that enterprise systems demand.
I’d be interested to hear what others in the NZOUG community think.
Where do you see AI having the biggest impact on your Oracle environment over the next three to five years? Will it be database administration, application development, enterprise integration, business applications, or something we haven’t anticipated yet?
Share your thoughts with the community, and let’s keep the conversation going.
Further reading:
Comments