It used to be just at Oracle OpenWorld, their annual show, that big advances in technology and applications were made. But now, with application updates pushed out every three months, announcements come much quicker, and so it was not unexpected that there would be new announcements at the AI World roadshow in London last week.
Oracle announced the AI Agent Applications—the next level in this fast‑paced AI revolution.
In just a few years, we have gone from redefining what is meant by AI, to the introduction of generative AI to help us create content, AI agents to help us do repetitive tasks, AI Studio to build your own, adding workflow, and now applying those agents to making decisions based on the outcomes we define.
Watch this interview interview with Steve Miranda from Bob Evans of Cloud Wars: . Bob always peels away the hype and quickly gets to the point—a powerful source of technology truth. Many others have posted their précis on LinkedIn, but I thought I would concentrate on what it means for users.
It took a while for people to get their heads around what was meant by AI, and I’ve previously said that I’ve started many a presentation with an AI explanation. But that’s no different from where we were 15 years ago when we were explaining what was meant by “cloud.” Generative AI almost crept up on us. ChatGPT was a phenomenon that frightened us—we thought it was going to take away every brain cell from our children and allow cheating in every possible academic scenario. But now we see it used for good (and sometimes bad) in our everyday lives. In many instances, it has become the default, and if there is an option, it’s usually to turn it off. In our applications, that’s slightly different. For Fusion Applications, a prerequisite was the move to Redwood, but once that profile option is flipped, generative AI is everywhere. I have a presentation where I talk about us at Inoapps using Fusion, and one place where we introduced this was in our goal‑setting process. There is good research showing that organizations using goals are more effective in supporting their employees, and it certainly helped speed up the goal‑setting process in our organization, where individuals and managers had access to generative AI.
The introduction of AI agents is where most people are now thinking about how they could use them in their organization. Recent commentary says that we’ve moved beyond proof of concept into execution, but I know that many organizations are still hesitant and still not sure where the value is for them. I read a really good article over the weekend, which talked about the most effective people having the most to lose with AI. If their processes are effective today, the risk of AI reducing that effectiveness is a bigger concern than for an organization that has a long way to go. It’s worth a read—change management is still a big consideration for your AI strategy.
Oracle quoted that 64% of Fusion customers are using AI, but in my discussions, that's any type of AI, not just AI agents. They also mentioned that those using agents were using less than 5. I think that number is important. AI for AI’s sake is not what people should be looking for—AI should be used to smooth out some of their challenges.
I have one customer who attended AI World in London who is currently looking at how AI agents can help them in their monthly payroll processes. The manual checking they do each month—no one wants to get payroll wrong. After Oracle announced the applications, we had a conversation about what that meant. In my mind, and remember I’d only had one day’s notice of the announcement, I was thinking of “applications” as being: How do you apply AI to your organization? How do you take the agents that can do the work that needs to be done and make a difference? So instead of telling your agents to check things at the end of each cycle, why not have them check things all the time and tell you when there’s a need to alter course?
I use the analogy of the Ever Given boat in the Suez Canal. When it left port it was on the right course, and its failure—the reason it became stuck—was that it didn’t alter that course as tides, winds, and life in general got in the way. If you’re using an AI application, it is looking at the outcome you need, the destination, and it can help you make those small course corrections.
I really do believe that AI is changing the world in which we live, and I’m honored to work in the technology that allows us to adopt it quickly and safely. Our data is in our organization, so many of the components an AI application needs to understand the organization are already in the application. Oracle’s strategy, as I define it, is the facilitation of AI wherever you need it. That means that in the application—and to use their words—it’s “built in, not bolted on,” and that means they can deliver it to us quicker than we can adopt it. More importantly, we can adapt as these exciting phenomena hurtle on at pace. We don’t have to build everything from scratch each time there's something we want to do.
There was also small change in the charging: AI is free within Fusion Applications, but if you extend AI agents or applications beyond Fusion, there is a small charge after a base allowance.
I have only talked about the Fusion Applications, but remember the technology that makes this work is in the Oracle technology stack, so it is available to build for any data.
