Wildbrush Collective

The client

Wildbrush Collective is an executive coaching and leadership consulting practice founded by Anna Smith (MHA, CPTM, ACC). It works with community health center, hospital, and health-system leaders, from newly promoted managers to executives building a leadership bench.

The practice needed a website that could do three jobs at once: feel as grounded and warm as the coaching itself, speak to two very different audiences, and turn interest into booked discovery calls without anyone touching a spreadsheet.

The challenge

• Two audiences with different needs: individual leaders looking for coaching, and organizations looking for programs and consulting.
• A calm, nature-rooted brand that had to feel premium and credible to healthcare executives, not like a template.
• Bookings, newsletter sign-ups, and enquiries had to land in the client's GoHighLevel CRM so follow-up happens automatically.
• A growing library of reflections, guides, and thought leadership that the client wanted to publish without a developer.

Design direction

The visual language borrows from the trail: an earthy palette of cream, rust, and bark; a serif display face paired with a clean sans-serif for body copy; painterly Montana landscapes; and hand-drawn brush-stroke buttons and botanical icons that mark each stage of the "Systems & Spirit" framework.

Every page keeps generous whitespace and a slow, editorial rhythm so the site reads like a conversation rather than a sales pitch.

Information architecture

The "two doors" pattern anchors the whole site. The hero, the first content section, and the contact section each offer the same two paths: "For You" leads to one-on-one coaching and a Trailhead discovery call, and "For Your Organization" leads to programs, consulting, and an organizational discovery call.

Beneath that, the experiences catalog presents Trailhead One-on-One Coaching, Rooted Manager Launch, Leadership Retreats, the Leadership AI Fluency Lab, and Organizational Consulting & Planning as consistent cards, each with its own landing page.

Build

• Next.js App Router with static generation for fast, crawlable pages and next/image for responsive artwork.
• Sanity CMS behind the Insights hub (reflections, guides, thought leadership) so the client publishes and edits content independently, with images served from the Sanity CDN.
• GoHighLevel integration through LeadConnector: two booking calendars (Trailhead and organizational discovery calls), a newsletter form with email and SMS consent, and enquiry flows that land in the CRM.
• A companion "Signal Check" tool on its own subdomain: an eight-question, five-minute AI-readiness assessment for healthcare leaders that ends with a personalized starting point and feeds interest back into the coaching offers.
• Technical SEO: metadata and Open Graph per page, ProfessionalService structured data with service areas, location pages, and a sitemap covering more than thirty pages.

Automation for the practice

Alongside the website, I built n8n workflows that take recurring manual work off the team's plate, and I am building a third that finds new business.

• A newsletter automation that pulls curated articles from RSS feeds on a schedule and drafts ready-to-send newsletter content for the practice, turning a manual research-and-write cycle into a quick review-and-send step.
• A specialized handbook generator that watches for completed client intake forms in GoHighLevel and triggers an n8n workflow to assemble a fully personalized client handbook automatically, removing a manual document-assembly task from every new engagement.
• A Healthcare Opportunity Radar, still in progress, that tracks healthcare organizations for consulting RFPs and calls for speakers. The next section covers it.

Lead generation: the Healthcare Opportunity Radar (in progress)

Consulting RFPs and calls for speakers are scattered across healthcare association sites, conference pages, procurement notices, and PDFs. Finding a relevant page is the easy part. Someone still has to confirm that submissions are open, find the right event cycle and deadline, and check the result against the existing tracker. The radar turns that research into a repeatable pipeline while leaving every pursuit decision with a person.

It is a 30-node workflow on self-hosted n8n. Google Sheets is the control surface: one tab for settings, one for the tracked sources, a 28-column Log of opportunities, and a Runs tab for processing counts and token usage.

• Fetches each tracked page or PDF, strips navigation and boilerplate, and fingerprints the cleaned text with SHA-256 so changed sources can be singled out.
• Sends readable pages to Claude for structured extraction that keeps consulting RFPs, dated speaker calls, and open rolling speaker forms apart. Missing facts stay "unstated" instead of being guessed.
• Validates outside the model in JavaScript: a deadline is accepted only when its supporting quote appears in the source, a year is never assumed, and event dates are never treated as submission deadlines.
• De-duplicates across runs and sources, records material changes such as a new deadline or link, prefers the issuing organization over aggregators, and never overwrites a reviewer's submitted or passed decision.
• Checks the estimated cost of every extraction request against a configured limit before sending it.
• Status: a fix that untangled the request-building and response-parsing stages passed 22 local checks with no live API calls or sheet writes. A controlled live run against known sources is the next acceptance step. Discovery accuracy and time saved have not been measured yet.