The Questions Enterprises Are Asking About Agentic AI Have Changed
Notes from a technical workshop in Jakarta on production-grade agent architecture

Yesterday in Jakarta, I engaged with some of the largest regional enterprises, addressing challenging technical questions on agentic architecture design, GenAI adoption, and modernization blockers. 🇮🇩 What I observed was remarkable.
The Shift I'm Seeing
Enterprises are no longer asking, 'What are the GenAI use cases?'
Instead, they're inquiring:
- 'How do my agent teams work across microservices?'
- 'How can agentic workflows and my team collaborate efficiently across projects?'
- 'What's the right autonomy boundary for production agents?'
The questions have evolved. The industry is prepared for this change.
AI-DLC Methodology + AWS Tooling: 4x - 7x Acceleration 🚀
The first deep technical takeaway: AI-DLC Methodology combined with AWS Kiro and Amazon Q Transform delivers 4x - 7x development cycle acceleration.
This is not about prompt engineering. It's a structured lifecycle that integrates code generation, autonomous testing, and continuous transformation. Participants transitioned from monoliths to agentic-ready microservices in just one afternoon.
The tooling has reached a critical point.
Decoupling Agent Orchestration from Business Logic
The standout architecture pattern was event-driven hooks that enable agents to reason over domain state without tight coupling.
This approach eliminates the 'brittle agent' anti-pattern - where a single schema change or API shift breaks your entire agentic workflow. It's essential for scaling collaboration between agentic workflows and human teams.
If your agents are tightly coupled to business logic, you're building technical debt into your autonomy layer.
Your Database Is Your Agent's Unlearned Knowledge Layer 🧠
This is the modernization many teams overlook.
Legacy schemas were not designed for autonomous retrieval and reasoning. It's crucial to restructure data models so agents can perform semantic queries, maintain context windows, and execute multi-step plans against the data layer - not just the API layer.
Every database is a knowledge layer that your agents have yet to learn to reason over. That's the unlock.
The Future of Enterprise Software Is Autonomy 💡
A massive shoutout to the team that delivered this incredible workshop. You made complex concepts accessible, and ignited real 'aha' moments in the room. 🙏
The future of enterprise software is not just about modernization - it's about autonomy. Every legacy system represents an agentic opportunity waiting to be unlocked.
I'm more excited than ever about the direction of agentic AI. The gap between 'demo' and 'production-grade' is closing faster than many realize.
If you're navigating app modernization and agentic AI, I'd love to hear about the challenges you're facing. Comment below or reach out - let's learn together.
#AlwaysDay1 #AgenticAI #GenAI #AIArchitecture #AppModernization #EnterpriseAI #AWS #WomenInTech
Gallery

The views and opinions expressed in this post are my own and do not necessarily reflect those of my employer or any organisation I am affiliated with.