Myles Gilsenan is the Vice President of Data, Analytics and AI at Apps Associates.
He’s a thought leader in AI, data strategy, modern data architecture and advanced analytics. He specializes in helping companies embrace and deploy AI at scale to transform into data- and AI-driven organizations.
We caught up with Myles to ask his opinion on Oracle’s new AI Data Platform.
In a recent blog, Myles made the point that one of the primary reasons AI initiatives fail is underestimating the need for clean, high-quality data. He notes, “Oracle AI Data Platform (AIDP) is quickly becoming the foundation for enterprise AI because it solves the data problem first.” AI succeeds when the data foundation is strong, and that’s exactly where this platform adds the most value, according to Myles.
He sees a number of use cases use cases emerging for AIDP and highlighted the 6 below.
- Unified data foundation: Bring structured and unstructured data together with governance, lineage, and security.
“When your boss says to you, ‘It’s time to get serious about AI. Our competitors are moving past us. If we don’t, we’re in trouble’, that means AI projects need to move past the POC stage and become full-fledged production AI applications. So, what’s the first thing you must do? Get control of your data. Not just databases from corporate systems of record, but ALL data — structured and unstructured: sales contracts, purchase contracts, policies, procedures, multimedia recordings, and images — should contribute to AI and machine learning applications. Oracle has cloud services for every one of those, and AIDP provides the place to consolidate all that data.
The data does not need to be exclusively in Oracle Cloud. Most enterprises have a multi-cloud policy, and AIDP acknowledges that. Whichever cloud holds your data, you can see all of it through the AIDP console. This is important because valuable data for AI initiatives comes from outside the enterprise: weather data, healthcare records, traffic patterns, customer behavior from marketing companies, and many more examples.”
- Faster AI and app development: Build AI-driven apps quickly with shared data assets, notebooks, and integrated GenAI services.
“Enterprises want to move quickly with AI, so they want to be able to use the tools, development frameworks, and models that their teams are most comfortable with. Some will want soup-to-nuts custom development with open-source solutions, others will want to use services from vendors like Oracle. But whatever approach is best, AIDP provides both the integration and the ability to manage the library of components for reuse.
When we first got hold of AIDP, we used it to build an application for forecasting construction project costs. It took us only a couple of weeks to put the entire thing together. We used open-source generative AI to populate data from historical projects, all held in PDF files. We created a data set to train our predictive model. The client wanted employees to be able to add their own data points using a front-end UI — ‘Hey, I’m thinking about a construction project. We’ve got about 150 units. It’s going to be 12 floors. It’s in Boise, Idaho’ — about 12 to 15 data points. With the AI application we built, they got predictions in five or six different cost categories.
With AIDP we can move quickly, reuse code, mix and match open source with Oracle Cloud services, and do it with full visibility and transparency. This enables us to go through the AI life cycle and track exactly what components we’ve put into production.”
- Intelligent analytics: Natural-language querying, automated insights, anomaly detection — all on governed data.
“We built an application to predict supplier risk using AIDP. We took in all sorts of data, including a lot from a third party, which we blended with some from Oracle Fusion Cloud ERP. In Oracle Analytics Cloud (OAC) we built a function to call the predictive model at runtime that gave them the full supplier risk score, using data about suppliers provided by users. AIDP enabled intelligent analytics because we were able to build the natural language queries, models, and agents into our analytics.
AIDP acts as the overall enabler and orchestrator for building intelligent analytics. It’s integrated with all the ingredients required for intelligent analytics — AI/ML model training and deployment, Oracle Autonomous AI Lakehouse, OAC and pipelines for AI agents.
Intelligent Analytics requires governed and curated data that can be trusted as a fundamental prerequisite for moving beyond just traditional, ‘what happened’ type analytics. One of the most popular architectural patterns for creating and serving up governed and curated data for analytics is the medallion architecture with its progressive curation of data across bronze (raw), silver (slightly transformed) and gold (fully curated) layers. AIDP natively supports the medallion architecture including the ability to access data in other clouds.”
- Scalable data engineering: High-volume ETL /ELT, feature engineering, and Spark pipelines in one platform.
“The compute layer of AIDP is Spark, and Spark is great for high volume processing and complex transformations. And if you’re used to a Python-based notebook experience for exploring your data, looking for anomalies, understanding the distribution of values, looking for things that have predictive value, you get that with AIDP. Python notebooks, always popular with data scientists, are becoming popular with many teams for doing traditional ‘ETl-type’ data engineering because they are scalable and flexible. There is a natural convergence taking place between traditional data engineering, machine learning feature engineering, and the pipelines required for AI Agents. Sometimes the same teams are doing all three and AIDP supports this type of collaboration and convergence.
Increasingly, streaming data is an important source, so AIDP integration with OCI Streaming and OCI Goldengate is useful. Maybe it’s sentiment analysis on a social media feed, or predictive maintenance data you’re getting from sensors, and you want to run a predictive model against it to act on the signals in real time. Or you want to capture and store predictions in an Oracle Autonomous AI Lakehouse to spot trends over time or share them in a dashboard or conversational interface for senior executives.”
- AI-powered automation: Predictive forecasting, exception handling, and embedded copilots across business processes.
“I would say automation is one of the most tangible and readily understood aspects of AI. For example, we no longer need to take purchase orders and manually figure out how to match them to the right invoice received as email PDF attachment. We can create an AI application to do that, and extract all the information from the invoice, put it into an ERP system, and tee it up for a human being to approve it. Or if the AI application determines it surpasses a particular confidence score, just have it automatically approved.
AIDP is ‘one stop shopping’ for creating AI Agents. It’s a central place and console where you can build complex end-to-end AI agents that incorporate all types of AI including predictive ML models. You don’t need multiple tools. Let’s imagine you are about to place an order and you’re dependent on the delivery date being met. In AIDP you could create a predictive model to assign risk ratings to suppliers and lead times for specific items. Then you could take that predictive ML and incorporate it into an AI Agent workflow to place orders based on supplier risk and lead time availability. When AI agents built in AIDP need to drive action in Oracle Fusion Applications, they can be exposed from AIDP as services or APIs and invoked by agents built in Oracle Fusion AI Agent Studio. In this way, AIDP-based agents and workflows complement application-native agents by providing deeper analytics and AI capabilities that extend beyond the boundaries of any single system. That’s phenomenal — other clouds don’t do that.”
As more and more agents are executing business processes, you’re going to want to know what’s out there, and what they’re doing, who owns them, what’s the risk level when they execute, etc. AIDP can catalog and document all the AI agents, as well as all the data sets being used in the enterprise.
- Governed collaboration: A shared catalog and consistent security so teams can work from the same trusted foundation.
“If you’re to have a broad-based, AI-led transformation, where it’s applied across every aspect of your company, in every department and function, it must be coordinated. And coordination starts with visibility. In a big organization it is often the case that everybody’s running their own initiatives on AI, creating the potential for a tremendous amount of redundancy. AIDP’s shared catalog is one of the most critical components for widespread AI success in any enterprise — because it provides visibility and the means to share and collaborate across multiple AI initiatives, promoting reuse, reducing redundancy, and boosting efficiency.”
AI is like oxygen
“AI is like oxygen — it’s everywhere, and it’s going to seep into every aspect of enterprises. It isn’t something that you can decide to do or not — AI is not optional. Is the internet optional? Are office productivity tools optional? No, of course not — they became pervasive and indispensable. AI will be the same.
What we’re hearing from clients in 2026 is that they want to stop dabbling and get serious about it. Most people at this point have had some exposure to AI and even to some advanced concepts of AI. You can go into a boardroom and say ‘RAG’ or ‘vector database’ or ‘medallion architecture’ and somebody will know what you’re talking about. It’s an inflection point and you can’t lurk around the edges — you have to dive in. And if you want to have success applying AI, you’re going to have to plan for it and manage it, no matter where in your company you’re rolling it out or who’s affected: your frontline staff, your back-office staff, the middle office, customers, products, etc.
Part of diving in is deciding which technology partners to work with. At Apps Associates we help our clients embrace AI as a transformative technology and scale AI across their enterprises. In our experience, Oracle Cloud applications and technology, coupled with Oracle AI Data Platform, form a powerful an end-to-end ecosystem and platform for AI-based transformation. we’re making those choices, both for our business and that of our clients. One of those choices is to adopt Oracle Cloud technology and Oracle AIDP in particular.”
Learn more about Myles and Apps Associates, and you can always ask questions in the Oracle Analytics Community.