AI Optimization
The practice, page by page
AI Optimization is making a business accessible, understandable and verifiable to AI systems, and proving it worked. These pages are the work itself, including the parts that argue against our own interests.
Being read correctly
What machines actually read on a page
Not what a person sees, and the gap is where most of the work is wasted or never started.
Consistency across sources, and why it does not average
Three descriptions that disagree lower confidence in all three.
Structured data: when it helps and when it hurts
A comprehension aid when accurate, a liability when it contradicts the page.
Being worth believing
What makes a claim verifiable
Checkable versus recordable, and what to do when the evidence does not exist.
Why publishing more is usually the wrong move
The reflex that most reliably fails, and the four cases where it is right.
AI optimization and SEO, in practice
Where they agree, where they diverge, and which tactics transfer badly.
Measuring it honestly
What a deterministic score can and cannot carry
Determinism says the ruler does not move. It says nothing about what you are measuring.
Signals decay, so a result is a statement about a date
What decays fastest, what holds longest, and the migration rule.
Owned-domain floors, and why we refuse them
Why our own properties score 0.5 on the largest measure.
What we refuse to score
An exclusion is where a thumb on the scale is easiest to hide, so the list is published.