Build UX intelligence

UX as a scalable operating system

TLDR

What was broken

Product knowledge was hard to trust across fast-moving systems and machine-readable surfaces.

What I did

I built clearer information structures, retrieval-readiness standards, and governance loops for product knowledge.

Why it mattered

Teams gained measurable AI-readiness, better source control, and stronger visibility into user-impacting changes.

80% Production Sources18 Changes Found22–26 Readiness

Philosophy

I design UX for humans and machines. As products become more conversational and more agent-assisted, UX leaders have to care about information structure, retrieval quality, and whether a system can explain itself clearly.

My point of view is simple: good UX is becoming part of AI infrastructure. If people cannot understand the product, machines will struggle too.

Strategy in Action

Retrieval-readiness framework

I defined how product knowledge should be structured so both people and AI systems can find the right answer with less ambiguity.

created a shared standard for clearer chunks, headings, and source quality

made retrieval quality visible as a product-experience concern, not just a technical one

Crawl-governance workflow

I created a cross-functional operating loop for identifying when automated systems were favoring temporary sources over canonical product information.

shifted 80% of automated crawls toward official production sources

connected source governance to customer trust and answer quality

Release-intelligence review

I established a repeatable way to surface user-impacting changes that might otherwise remain invisible across fast-moving product teams.

surfaced 18 untracked changes in three weeks

turned hidden maintenance risk into an observable operating signal

Content-quality signals

I used lightweight automation to make unclear language, missing context, and maintenance debt easier to spot before they became customer friction.

reduced manual review toil

helped teams act on clarity issues earlier in the workflow

Outcomes: clearer product knowledge for people and machines

Created the conditions for teams to treat AI-readiness as product experience work, not a side project.

Outcomes included 80% of AI crawls reaching official production sources, 18 untracked user-impacting changes surfaced in three weeks, and a measurable retrieval-readiness score reaching 22-26 points for production-ready materials.

AI-ready product knowledge

Improve the source, improve the answer.

Find

Identify where people and systems are getting product truth.

Structure

Make each section clear, complete, and independently useful.

Govern

Route systems toward canonical sources and away from stale ones.

Measure

Track retrieval quality, source accuracy, and answer clarity.