I build enterprise AI — and defend the business case for it
Sixteen years across banking, telecom, energy and government in Southeast Asia. Hands on the architecture and the delivery, and in the room when the money gets approved.
- 16+ years
- $15M+ ARR driven
- $30M+ AI value realized
Open to speaking, consulting, and ghostwriting — hello@satsawat.ai
The Builder in the Boardroom
I'm Satsawat Natakarnkitkul — a Data & AI Leader for ASEAN. For 16 years I've been building enterprise AI systems across banking, telecom, retail, energy, and government in Southeast Asia.
I do both halves of the job. I build — architecture, pipelines, models, the debugging and the unglamorous data work. And I sit with C-level on strategy, business cases, and what the investment actually returns.
Most people in this space are one or the other: strategists who've never shipped a system, or engineers who've never had to defend a number to a CFO. The gap between them is where most enterprise AI quietly dies. That gap is what I write about.
This site is where I share architecture patterns, strategy frameworks, and honest lessons from the field.
I'm also the author of two books on agentic AI in production and what AI means for your career — see the Books section below.
Interested in working together? I'm open to speaking engagements, consulting, and ghostwriting for executives and organizations.
All opinions and views expressed on this site are my own and do not represent the views, positions, or policies of any current or former employer, client, or organization I am or have been affiliated with.
Recent Writing
Architecture teardowns, strategy playbooks, and honest perspectives on enterprise AI.
The Semantic Layer Is the New API for AI Agents
AI agents hallucinate because your data has no shared vocabulary — not because models are bad. The semantic layer fixes that for humans and machines alike.
Read article →Why Your AI Agent Needs a Performance Review (Literally)
An agent that resolved 85% of requests in testing saw customer satisfaction drop 12% within three weeks of production. Your logs showed green across the board. Traditional monitoring missed it entirely.
Read article →Your AI Agent Just Made a Decision. Who's Accountable?
80% of organizations have encountered risky behavior from AI agents. Three frameworks are emerging to fix this — but most enterprises haven't adopted any of them.
Read article →What I Write About
The six areas I write about most. Each one links to everything I've published on it.
Agentic AI in Enterprise
Multi-agent architecture, governance, failure modes, and the path from pilot to production.
Read these →AI Architecture Patterns
Real-time inference, API-first ML platforms, RAG patterns, and systems that serve millions.
Read these →AI Strategy & ROI
Measuring real AI value, building business cases, and what actually generates revenue.
Read these →AI in ASEAN
Adoption patterns, regulatory landscape, regional case studies, and opportunities across Southeast Asia.
Read these →Data Science Craft
From notebook to production — ML engineering, feature engineering, and analytics at scale.
Read these →Data as the Foundation
Bad data in, bad AI out. Data quality, governance, pipelines, and why your AI is only as good as what you feed it.
Read these →Long-Form Work
Two books on the gap between AI's promise and what actually ships — written by someone who's built the systems, not just watched them.
Why Your AI Agent Will Fail
And What the 11% Who Ship to Production Do Differently
Only 11% of enterprise AI agent pilots reach production. A practitioner’s playbook for the decisions — not the demos — that separate the two.
REPLACED
Why I Built the AI That’s Restructuring Your Career and How to Stay Ahead of It
You’ve been good at your job for years. Now AI does the 80% you’ve been running on autopilot — a framework for finding where your career value actually lives.