Designed & built Case study In development

USA Archery Platform Modernization

A working modernization of a national governing body's WordPress site — 2,900+ articles migrated to Payload CMS, a rebuilt information architecture, role-based editorial workflow, and a governed AI assistant with citation-backed answers.

Project
USA Archery (proposal & working demo)
Role
Principal — strategy, IA, and implementation
Timeline
2026
usa-archery.vercel.app
USA Archery demo homepage with the headline Archery Is For Everyone over a photo of archers, primary navigation, and the Ask Nock assistant button
2,900+
Articles migrated
739
Resources catalogued
~1-2 min
Build time, from 6+

A national governing body’s platform, rebuilt

USA Archery carries a lot on one website: Olympic and Paralympic team information, membership, coach and judge certification, event sanctioning, SafeSport compliance, and a large archive of member resources. All of it ran on a WordPress installation that had grown difficult to manage and expensive to maintain.

I built the modernization as a working demo, deployed and explorable, running the organization’s real content.

Organizing around who actually arrives

The legacy site buried everything in deep sidebar navigation, and audiences with completely different goals shared the same paths. A parent checking SafeSport and a coach seeking certification were treated as the same visitor.

I audited the legacy navigation and rebuilt the information architecture around the five people who actually show up: someone who wants to try archery, a member renewing, a coach seeking certification, a parent checking SafeSport, and a fan following the national team. The deep sidebar became a mega menu that shows the full depth in the open, and every page carries a contextual sidebar so visitors can see where they are and what sits alongside it.

I documented the work as artifacts the organization keeps: an information architecture map, decision records for the platform choice and the migration, a design system, and a content-ownership matrix naming exactly which content staff control and which developers do.

Twenty years of content, made findable

2,900+ articles, 134 pages and 739 downloadable resources moved to the new platform, cleaned up on the way — broken legacy links repaired, thin pages routed to review rather than published, and every resource sorted into one of 21 categories.

Moving an archive is only worth doing if people can find things in it afterwards. News, resources and site search all share one filtering vocabulary, so a search for “coach” narrows 2,903 articles to 264 and keeps narrowing by category. Visitors can filter 207 national team athletes by division and discipline, and 41 events by state and by competition level.

Nock: KCS reasoning as conversational AI

The assistant is named Nock, after the notch on an arrow — the first step before every shot. It answers in three tiers, and the ordering is the design:

  1. Curated answers. Staff write the answers to the questions they get most, and those come back instantly, worded the way the organization wants.
  2. Answers from published content. With no curated match, Nock searches the site’s own pages and articles and answers from them, showing which sources it used.
  3. Model fallback. Reached when the organization’s own content has nothing to offer.

This is KCS structure expressed as conversational AI. Curated answers are the known-good articles; cited retrieval is search against the knowledge base. The organization’s own words come first.

Governance matters as much as retrieval. Nock answers from what USA Archery has published, routes SafeSport questions to the U.S. Center for SafeSport, keeps athlete information private, and says plainly when a question falls outside what it knows. Staff manage the curated answers themselves, and the panel tracks how often each one is used — so the questions that fall through to the model become a measured list of content worth writing.

Built for the way the organization works

Staff publish through the CMS the way they would on any content platform, with no developer involved. The site is tuned for a governing body’s traffic shape, where an Olympic medal can arrive as a sudden spike.

Editorial permissions mirror what the organization already does in WordPress: five roles, with authors drafting and editing their own news and events, editors and above updating anything, and membership records restricted to administrators. Logged-in editors get an edit button on the page itself, linking straight to the record behind it.

Design follows the organization’s published brand guide — Poppins, and the specified Pantone-matched navy and red.

Inside the build

5
usa-archery.vercel.app
Participate mega menu expanded into three grouped columns — Get Started, Youth & Schools, and Disciplines — beside a Find a Club Near You promo card
The rebuilt IA: a mega menu that carries the legacy sidebar's depth without hiding it, grouped by what the visitor came to do.
usa-archery.vercel.app
News listing showing 12 of 2,903 articles with category filter chips and a year-by-year archive sidebar
2,903 migrated articles, made findable: keyword search, category filters, and a year archive with per-year counts.
usa-archery.vercel.app
Search results for the query coach, reading Showing 12 of 264 results, with facet chips for Competition, Team USA, Youth and JOAD, Olympic and Paralympic, Coaching, Collegiate, Club News, Awards and Recognition, and Events
Search across the migrated archive: 264 of 2,903 articles match "coach", narrowed by the same category facets used on the news listing.
usa-archery.vercel.app
Resource Center with a search field, Most Requested PDF cards showing categories and file sizes, and a row of category filter chips
739 documents catalogued — 652 auto-categorized by keyword matching, the rest left for human review.
usa-archery.vercel.app
The Nock assistant panel answering how to become a member with a structured breakdown of membership tiers, pricing, and benefits, plus follow-up question chips
Nock answering from a staff-authored curated intent — exact, instant, and worded the way the organization wants. Retrieval answers add a 'Based on' row of source titles; the model fallback is last.

Questions about this work?

Happy to go deeper on any of it — the decisions, the trade-offs, or what I'd do differently.