We run our own playbook on this website.

We decided ours should. If we ask a bank or credit union to structure content for AI, add schema, welcome AI crawlers, and fix accessibility, the least we can do is hold our own site to the same standard and let you inspect it.

Every technique below is live on the page you are reading. Open the source, follow the links, and run your own scan. The method is the same one behind a benchmark that has audited 230+ financial institutions.

The core of whether an AI engine can parse, trust, and cite you.

What we do here: a single entity graph, one Organization node and one Person node, that every page references by ID, so an assistant reads a consistent picture of who we are. FAQ schema on the pages where people ask real questions. An answer-first summary near the top of key pages, written so an assistant can lift it cleanly in a sentence or two. Speakable markup on the parts meant to be read aloud. Semantic HTML landmarks. And an llms.txt file that indexes the site for agents in plain language.

Check it yourself: view this page's source and read the JSON-LD blocks in the head, then open llms.txt.

Still the foundation. The signals that rank a page also help an AI engine trust it.

What we do here: a unique title and meta description on every page, one h1 and a clean heading hierarchy, canonical URLs so nothing competes with itself, descriptive internal links between related pages instead of generic "click here," and a complete XML sitemap kept current as pages change.

Check it yourself: open sitemap.xml, and read any page's title and headings in the source.

A slow or broken site signals that an institution lacks the maturity to recommend. Performance is a trust proxy.

What we do here: static HTML that loads fast with no framework overhead, HTTPS everywhere, a robots.txt that explicitly welcomes AI answer engines instead of blocking them, cache headers tuned per file type so updates propagate correctly, and a layout that holds up on mobile where most financial search now happens.

Check it yourself: open robots.txt and read the welcomed crawlers, then run any page-speed test against the domain.

Accessible HTML is also the structure AI parses best, so this work pays twice. And in financial services it lowers real legal risk.

What we do here: a skip-to-content link, semantic landmarks for navigation, main, and footer, alt text on images, color contrast that meets WCAG AA, menus you can operate from the keyboard, and forms with real labels rather than placeholder text standing in for them.

Check it yourself: tab through this page with your keyboard, and run any accessibility checker against the URL.

The honest one. Most reputation signals live off your own site, in reviews, press, and citations you do not fully control.

What we do here: what we control, we keep clean. A consistent name, address, and contact point everywhere they appear. A named and linked author on our work rather than an anonymous byline. sameAs links to our real profiles so an engine can connect the dots. And clear disclaimers wherever we cite third-party data, so a reader and an AI engine can both tell what is ours and what is sourced.

Check it yourself: compare the contact details in the footer, on the contact page, and in the Organization schema. They match on purpose.