If you missed Oracle AI World (formerly Cloud World) last week, you missed a firehose of information. With 20,000 delegates and 900 sessions, it’s a lot to digest. We tuned into the keynote from CTO Larry Ellison to cut through the noise and understand Oracle’s grand strategy.
From a marketing perspective, it’s clear: Oracle is making a massive, all-in gamble on AI. But their approach is different from competitors like Salesforce, and it hinges on one central, trillion-dollar conundrum.
Here are our key takeaways.
LAST UPDATED: 28th September 2025
AI isn’t hype, it’s already delivering
Before the main keynote, the most compelling stories came from real-world “AI in action” ads:
- Slashing Hiring Time: One consulting firm is using AI in HR (an Oracle strong suit) to vet employees, analyse interviews, and summarise notes. The result? They’ve cut their time-to-hire from 45 days down to just 21 days.
- Streamlining Travel: US Customs at JFK Airport is using AI-powered facial recognition to identify US citizens, allowing them to walk straight through customs without stopping.
These examples show that AI is moving from a theoretical “what if” to a practical tool for massive efficiency gains.
*SOURCE: CNBC | OpenAI’s first data center in $500 billion Stargate project is open in Texas, with sites coming in New Mexico and Ohio | Tuesday 23rd September, 2025
Oracle is building this huge data centre. It’s a 1.2 billion-watt AI brain that has enough power for 4,000 four-bedroom homes. It’s interesting that it takes such a huge amount of power to process all of this data in real time and get the answers back that people want.
Oracle’s massive share price increase is due to its backing to build these huge data centres and, essentially, power this AI revolution. The challenge is that the revenue won’t come until after they’ve built all these data centres.
Adrian Jones | CEO & Founder | Marketing Cube
The trillion-dollar conundrum: Using private data (While keeping it private)
Here’s the most interesting part of the strategy. Oracle’s databases hold most of the world’s highest-value private data—think health records, government data, and tax information.
The problem? Customers want the benefits of AI’s reasoning, but they do not want their private data to be used to train AI models.
Oracle’s solution is to separate the concepts. Their strategy is to use private data for reasoning, not for training.
How? Through technology like Retrieval-Augmented Generation (RAG) and AI vector search. Think of it this way: the AI can learn that “hot water burns your hand” (a reasoning-based conclusion) without ever needing to know the specific data point of “what temperature the water was” (the private data).
This allows Oracle to build smarter models by learning from the patterns in private data without ever seeing, sharing, or compromising the data itself.
What this actually means for marketers: The new agents
Based on discussions with the Oracle team, Oracle plans to roll these agents (and others) out in 2026. If you subscribe to our Important Eloqua Updates service, delivered quarterly, you’ll be among the first to know.
Account Product Fit Agent
This agent analyses your account data and engagement signals to predict which customers are most likely to buy another product and which product it will be. It’s a predictive ICP (Ideal Customer Profile) model on steroids.
Buying Group Definition Agent
This is a huge one for personalisation. Most marketing teams can barely manage 2-3 personas. This agent is designed to help marketers effectively personalise strategies across 5, 6, or more personas by aligning content with each persona’s specific journey stage.
Model Qualification Agent
This is a more advanced segmentation tool, using predictive modelling to define your “best-fit” audience for campaigns more accurately.
Our two big takeaways
This all sounds great, but the presentation ended on two critical action points for any business (including us) looking to capitalise on this revolution.
- Get Your Data House in Order. AI agents are useless if your data is a mess. If the data’s meaning isn’t obvious, an AI won’t be able to help. The #1 focus for every company right now must be data quality and governance.
- Upskill Your Team to Become “AI Coaches.” The goal isn’t to be replaced by AI; it’s to work with it. We need to shift our mindset to become “AI collaborators,” learning to ask the right questions and guide the technology to achieve the best results.
AI is here to stay, but we are definitely in the “hype curve.” A dip will come when people realise the challenges of preparing their data. But after that, the businesses that invested in clean data and skilled “AI coaches” will be the ones who truly pull ahead.
Hero photo by Leo_Visions on Unsplash | The Venetian Conference & Expo Center, Las Vegas, NV, USA
All other images in this blog post have been generated by Marketing Cube using Google Gemini.
