Build Your Agent Stack
Tools, models, infrastructure, and the difference between chatbots and agents.
THE PRACTICAL GUIDE TO AI AGENTS
AI agents are leverage, not magic.
The Agent Advantage is a field guide to using AI agents to build, automate, and operate real businesses—without handing away your judgment or believing the hype.
Written by an engineer who uses AI agents to run stores, build software, test ideas, and automate real work.
THE WALL CAME DOWN
What changes when the work you hate is no longer the barrier?
In 2007, Alastair built a fantasy-football website by hand. It worked. Almost nobody came—and the deeper failure was how much he hated the tedious programming required to test the idea.
For fifteen years, ideas stopped between concept and execution. Agents shortened that distance. They did not make every idea good. They made ideas faster and less painful to test.
Read the full story →
INSIDE THE BOOK
A working system, not a promise.
From choosing a stack to putting hard limits around what it can do.
Work With Agents Effectively
Research, propose, approve, refine, verify—and keep human judgment in control.
Apply Agents to Real Businesses
Stores, content, websites, software, digital products, marketing, and advertising.
Build Guardrails Before Scale
Execution ceilings, spend caps, monitoring, approval queues, permissions, and kill switches.
Launch One Useful System
A 30-day plan: learn manually, then automate one station at a time.
THE LINE IN THE SAND
“If you want magic, this is the wrong book. If you want leverage, start here.”
For the person willing to supervise
Side-business operators, solo founders, creators, print-on-demand sellers, non-programmers, and technical professionals who will verify the work.
Not a passive-income fantasy
No guaranteed revenue. No fully autonomous business. No machine that decides what is worth building. No automating a process you have never done.
MEET ALASTAIR FRASER
An engineer’s approach to AI agents
For most of his career, Alastair Fraser has worked on systems where requirements, testing, failure modes, and human responsibility matter. He brings that same mindset to AI agents.
BUILT, BROKEN, IMPROVED
Projects as evidence
ATS Resume Designer
A working core appeared in a day. Real-world files exposed the gap between a demo and something trustworthy. Roughly 100 messy résumés, more than 20 fix-test cycles, and a growing regression suite did the real work.
Read the case study →
Agentic Bot Sitter
A reviews-and-news site built around evidence, separate editorial skills, and a rule that the system asks rather than invents when a fact is missing.
Read the case study →
Agent-assisted business systems
Online business workflows for research, designs, listings, captions, and marketing—kept behind human inspection, platform rules, and approval points.
Read the case study →OPERATING PRINCIPLES
Rules written after the failure.
- 01Do it manually before you automate it.
- 02Agents propose. Humans decide.
- 03Automate one station at a time.
- 04Put limits in code, not just in policy.
- 05Measure behavior, not compliments.
- 06Keep a kill switch within reach.
- 07Use deterministic code for exact or irreversible work.
- 08The model is rarely the first bottleneck.
PROFESSIONAL WORK
Building serious systems with emerging tools
Alastair is interested in thoughtful work involving AI agents, systems engineering, applied automation, architecture, prototyping, hardware/software integration, technical leadership, and human-in-the-loop systems.
THE AGENT ADVANTAGE
A Practical Guide to Using AI Agents to Build, Automate, and Run Your Business
What actually works, what breaks, and what the hype leaves out.
Get the Book on AmazonAs an Amazon Associate, I may earn from qualifying purchases.
The hard part is no longer building everything yourself. The hard part is deciding what deserves to be built—and keeping your judgment when an agent helps build it.