Experiment
Does AI-Readable Portfolio Architecture Increase Digital Asset Leverage?
A live experiment testing whether a portfolio-style site with explicit framework pages, role-based content types, AI-facing files, and cross-asset logic becomes easier for machines and users to classify, trust, and route value through.
If a digital asset property publishes clear framework steps, named asset classes, explicit portfolio relationships, and AI-facing orientation files, then it should become more legible as a source, easier to navigate as a system, and more valuable as a leverage asset than a comparable site that presents the same ideas only as generic marketing pages.
What is being tracked.
How the test is structured.
The experiment uses LeverageBuilder.com as the live subject. The baseline observation point is June 6, 2026, after the initial framework, asset-class, portfolio, tool, about-page, `llm.txt`, and `ai/manifest.json` layers were documented in the repo.
At baseline, the project already defines six asset classes and five framework steps. It also publishes AI-facing files that describe the site's topics, founder expertise, content types, and citation preference. The core question is whether that structural clarity creates measurable leverage signals over time.
The observation model is not limited to rankings. It tracks whether system architecture improves classification and routing across five dimensions: page-type indexation, branded discovery, AI citation or mention behavior, internal linking between framework and portfolio layers, and assisted conversion behavior where more than one page type participates in the visit.
The practical comparison point is the standard strategy site that explains digital growth in prose but does not expose named system parts, explicit relationships, or machine-readable orientation files.
What has happened so far.
September 11, 2026 checkpoint.
The structural layer is still the strongest result. LeverageBuilder has framework pages, six asset classes, tools, portfolio records, and machine-readable orientation files.
Search exposure expanded. June produced 53 impressions, 8 clicks, and visibility across 3 pages. The latest 28-day window produced 694 impressions, 1 click, and visibility across 20 pages. The same recent window recorded 171 successful crawler content requests.
That is broader discovery, not a conversion win. Clicks did not grow with impressions. There is still no isolated control, confirmed AI-citation record, or assisted-conversion measurement showing that the machine-readable architecture caused the expansion.
What would support or challenge it.
Status: open, with exposure growth and no leverage proof yet.
The architecture is more legible and more of the site is appearing in Search Console. That is worth recording. It is not enough to claim traffic lift, lead lift, citation lift, or valuation lift.
The next useful measurement is no longer another page count. It is whether named framework and asset-class pages earn non-branded discovery, citations to specific concepts, multi-page human sessions, or qualified inquiries that can be traced through the system.
Keep Following the Tests
Move from this open thread back into the full experiment library.