From CRUD to Cognitive: Modernizing Legacy Java Apps with Embedded AI

Transform your legacy Java apps into AI-powered systems without starting from scratch. From CRUD to Cognitive is your guide to a smarter, AI-driven future for enterprise apps! Join Ed Burns (Principal Architect for Java on Azure) to learn how to bring intelligence into traditional Java applications. Tasked with a "do something with AI" mandate? This session shows Java enthusiasts how to embed powerful GenAI capabilities into your apps. Using open-source frameworks and models, you'll learn to enhance workflows with Retrieval-Augmented Generation (RAG), implement AI tools like form auto-fillers, streamline operations with smart chatbots, and optimize workflows with AI triggers—all without ripping apart your existing systems.

Transcript

1. From CRUD to Cognitive: Modernizing Legacy Java Apps with Embedded AI Ed Burns,Principal Architect Microsoft
2. Ed Burns professional biography  Client  NCSA Mosaic (1994)  SGI Cosmo Web Authoring  Sun Netscape 6 OJI  Server  J2EE JSF (2002)  Oracle Java EE  Servlet, JSF, Bean Validation, etc.  Cloud  Java Champion  Microsoft Azure Cloud (2019)  Jakarta EE 11 committer and release coordinator  Books
3. My plan for your time investment  Clarity is emerging  skill at being an “agent boss” offers best odds for maintaining economic utility.  Idea book for adding AI to legacy app.  Hacking on Cargo Tracker  RAG  AI Shortest path
4. Best odds for remaining economically relevant: Skill at being an agent boss
5. Every team will achieve more with AI and agents Customer Service & Support From To Sales & Marketing Finance & Operations DevsBusiness (that’s us) 82% 1Bn New logical apps forecasted to be created by 20282 Supply Chain Business leaders expect agents to be moderately or extensively integrated into their company’s AI strategy in the next 12–18 months.1 Human Resources Manual, broad-based engagement JudgedStatic on reporting the code Fragmented, reactive Generic process and planning logistics and surveys you personally write, A Frontier Firm is a company thatwork leverages during hours LLM-powered AI agents alongside human employees automate workflows, scale on demand, and Work Trend Indexto Annual Report: 2025: The Yearcapacity the Frontier Firm is Born operate with a new organizational structure where every worker directs and https://aka.ms/2025-work-trend-report manages AI agents. Judged on the Real-time, Intelligent, Personalized, outcomes your Predictive, Hyper-personalized, AI-driven decision adaptive AI-empowered autonomous support AI-augmented selling agents deliver, making 24/7 orchestration journeys 365 Reactive issue resolution 1. Work Trend Index Annual Report: 2025: The Year the Frontier Firm is Born 2. https://www.idc.com/getdoc.jsp?containerId=US51953724 B
6. Work trend report: digital labor Organizations will need to consider if there are times when human and digital labor outperform AI alone, when customers prefer companies find itain human their interest touch orto when deliver society a human expects touch people or whentosociety be responsible expects people for the to consequences be responsible — for likethe a high-stakes product consequences — likeorafinance high-stakes decision. product or finance decision. To maximize the impact of these human-agent teams, organizations need a new metric: the human-agent ratio. Leaders must ask two critical questions: How many agents are needed for which roles and tasks?
7. https://bit.ly/susskind-ai
8. https://bit.ly/susskind-ai vs.
9. What will remain for developers to do? Best odds strategy • Maximize your proficiency in being an agent boss. • Right now, “the ralph loop” is hot. • Build skill at making it so you can keep agents busy doing productive things while you are not working.
10. Let’s get to adding AI to legacy apps
11. GitHub Copilot modernization Now with the modernization agent! Modernize at scale Use the modernization agent to assess applications at scale, plan application specific journeys, and execute with automated upgrades Customize and Repeat Tasks Use custom skills to encode your own business logic, software factory patterns, and standardize outcomes Automate Java and .NET upgrades Upgrade language runtime versions and frameworks effortlessly with AI-driven remediation and automated error fixes Streamline end-to-end migration From assessment to deployment, simplify the migration journey with full visibility, control, and predictability, while delivering higher quality outcomes. Embedded where developers work Accessible in GHCP CLI, CCA, Modernization CLI, VS Code, Visual Studio, IntelliJ https://aka.ms/ghcp-modernization
12. Develop 6R strategy on an application level Start No Application still aligned Yes with business goals (i.e., no massive changes required)? No Retire Business-relevant Yes application? No Rearchitect or Rebuild Does application need code Yes level or architectural level changes? Yes Replace Modernize Replatform Can workload leverage Migrate No PaaS with config level changes? Rehost E
13. Adding AI to legacy apps Idea book AIAI
14. Low hanging fruit Search • Replace or augment clunky keyword search with semantic search — users describe what they want in natural language instead of guessing field names or exact terms Summarization • Long text fields (notes, comments, descriptions) get autosummarized, useful in list views where you can't show everything Data entry assistance • User starts typing into a form, LLM suggests completions or flags inconsistencies before submission
15. Reducing the pain of legacy UI Natural language querying • Let users ask "show me all orders over $10,000 from last quarter that haven't shipped" instead of navigating complex filter screens — the LLM translates intent into the query your existing data layer already knows how to run Form navigation assistant • A chat interface that guides users through complex multi-step forms, replacing the need for a manual or training
16. Concrete examples Hacking on Cargo Tracker
18. Demo 1: RAG informed AI chatbot for tracking https://github.com/azure-javaee/2026-javaland-crud-to-cognitive cargotracker-rag B
19. Demo 2: Use AI to compute shortest path https://github.com/azure-javaee/2026-javaland-crud-to-cognitive cargotracker-ai-shortest-path B
20. Next steps 1 Assessment Assess your estate & plan your migration and modernization journey with tooling, guidance, and recommendations 2 Pilot & proof of value Gain confidence and see technical proof with expert support from Microsoft and partners 3 Programs and guidance Securely move your estate to the cloud with confidence, proven best practices & funding to accelerate Invent with purpose B
21. From CRUD to Cognitive: Modernizing Legacy Java Apps with Embedded AI Ed Burns,Principal Architect Microsoft https://www.linkedin.com/in/edburns/