Peter’s Stack is a technology blog about Enterprise AI, automation, CRM, and SaaS, viewed through the lens of real business processes and system design.
Rather than listing features of the latest tools, the site focuses on what role a technology should play inside a real enterprise environment and how it should connect with existing systems.
About Peter
I spent roughly a decade in B2B manufacturing sales, working with dealers, OEMs, and customers across Korea and international markets on long-cycle sales and projects. That experience taught me that business outcomes are shaped not by technology alone, but by the way customer work, decision-making structures, systems, and people fit together.
Today, I work in the Enterprise CRM and AI transformation space, with a focus on architectures and workflows that connect AI with existing business systems. My interests include CRM platforms such as Salesforce, LLMs and RAG, agentic workflows, business automation, and API-based system integration designed for real operations.
What Peter’s Stack Covers
- Enterprise AI and Agentic Systems — deciding what AI should be allowed to reason about and what should remain under deterministic system control
- RAG and Knowledge Architecture — turning enterprise documents and tacit knowledge into searchable knowledge that LLMs can use reliably
- CRM and Workflow Automation — connecting AI with existing business systems, including Salesforce, without disrupting established operational boundaries
- AI PoCs and Runtime Design — using tools such as n8n, APIs, LLMs, and coding agents to validate ideas quickly and evolve them toward operable systems
- Enterprise Architecture — permissions, state management, observability, cost, and failure handling that are easy to overlook during AI adoption
How Articles Are Developed
Articles on Peter’s Stack start, wherever possible, from problems I have worked through directly, implementation experiments, and architecture decisions. When discussing products or technologies, I also refer to official documentation, public research, and other primary materials, while distinguishing those sources from my own observations.
The goal is not to document only successful outcomes. I also try to preserve the original assumptions, failed approaches, trade-offs, and limitations that make an implementation useful to other practitioners. Confidential company or customer information is not published; when real-world examples are needed, they are generalized or reconstructed within a publicly shareable scope.
What You Can Expect Here
Peter’s Stack is not intended to be a stream of AI news summaries or unconditional tool recommendations. The recurring question is: “What has to change for this technology to work inside a real business?” The aim is to build a reusable record of architecture choices, implementation patterns, and lessons from practice.
For questions about the site or its content, visit the Contact page. Information about privacy, cookies, and site data practices is available in the Privacy Policy.
Peter