Application Modernization and AI Services
Domain 5 - Application Modernization and AI Services. This domain tests choosing a modernization path and using IBM AI services responsibly, including when a managed AI service beats building a model.
Modernization paths
The 6 Rs are retire, retain, rehost, replatform, refactor, and repurchase. Rehost moves as-is; replatform changes runtime or middleware; refactor changes architecture, often to microservices. Containers package apps and dependencies; Kubernetes or OpenShift orchestrates them; Code Engine runs serverless containers, jobs, and functions. CI/CD toolchains automate build, test, scan, and deploy. Modernization should be driven by business capability, not by containerizing everything.
AI services
watsonx.ai provides foundation models, prompt engineering, tuning, and deployment; watsonx.data provides a governed lakehouse; watsonx.governance manages risk and model lifecycle. Watson Assistant builds conversational experiences; Watson Discovery searches enterprise content; Speech to Text, Text to Speech, and Language Translator add modalities. Retrieval-augmented generation grounds answers in enterprise data and reduces hallucination.
Worked example
A call-center team wants answers from policy PDFs with citations. A custom foundation model is unnecessary. Ingest PDFs into a governed data layer, chunk and index them, use watsonx.ai with retrieval-augmented generation, and add watsonx.governance for evaluation and audit. Watson Assistant can provide the conversational front end.
Traps
Lift-and-shift is not modernization; a foundation model is not automatically compliant or accurate; RAG does not eliminate the need for data quality and access control; and serverless is a poor fit for long-running stateful workloads.