Why AI?
Why AI with IBM i?
Section titled “Why AI with IBM i?”IBM i is home to decades of mission-critical business data — financial transactions, inventory records, customer history, and operational telemetry — accumulated across industries like manufacturing, distribution, retail, and banking. That data is already clean, reliable, and stored in Db2 for i, one of the most battle-tested relational databases in existence.
AI works best when it has access to high-quality, well-structured data. IBM i already has that. The question is no longer whether AI makes sense with IBM i — it’s which AI approaches unlock the most value from what you already have.
What AI can do for IBM i workloads
Section titled “What AI can do for IBM i workloads”Extract insight from existing data
Section titled “Extract insight from existing data”Classic machine learning techniques (regression, classification, clustering) can be applied directly to Db2 for i data to uncover patterns that were previously invisible — detecting anomalies in transactions, predicting equipment failures before they happen, or segmenting customers based on purchasing behavior.
Augment existing applications with intelligence
Section titled “Augment existing applications with intelligence”Rather than replacing existing RPG, COBOL, or CL applications, AI can be layered on top of them. An existing order-entry application can be augmented with a demand-forecasting model; an inventory system can gain an AI-powered reorder recommender — all without rewriting the core business logic.
Enable natural language interaction
Section titled “Enable natural language interaction”Generative AI and large language models (LLMs) make it possible for users to ask questions of IBM i data in plain English, generate SQL queries from natural language, or build conversational interfaces over existing data and workflows.
Automate and orchestrate complex tasks
Section titled “Automate and orchestrate complex tasks”Agentic AI — AI systems that can plan, reason, and use tools — can interact with IBM i through APIs and SQL, orchestrating multi-step workflows that previously required human intervention.
Why keep AI close to the data?
Section titled “Why keep AI close to the data?”Moving large volumes of data to a cloud AI platform for every inference request introduces latency, cost, and security risk. Running AI close to — or directly on — IBM i keeps sensitive data on-premises, reduces round-trip time, and takes advantage of IBM Power’s hardware acceleration capabilities (MMA on Power10, Spyre accelerator) to run inference efficiently.
Where to start
Section titled “Where to start”If you’re new to AI on IBM i, the Choosing an AI Stack page walks through the landscape of tools and platforms available, from IBM-native solutions to open-source frameworks. The Use Cases section covers specific journeys: Machine Learning, Deep Learning, Generative AI, Agentic AI, and Developer Experience tooling.