Boulder SEO Marketing - AI search optimization toolkit

Challenge

Boulder SEO Marketing built its reputation on the Micro SEO℠ method: find a page ranked 2 or 3, lift it to position 1, and simultaneously earn citations in generative search (ChatGPT, Gemini, Claude, Perplexity). The methodology worked. The problem: it was delivered by hand, in Google Sheets, by one founder. Not scalable

They needed internal tooling that turned the playbook into software without losing the "founder-level" judgment their clients paid for

Discovery

We shadowed the founder on a client audit and documented every decision point: why pick this keyword, this page, this rewrite angle? Most decisions came down to three signals: search intent match, existing page relevance score, and whether AI engines were already citing competitors. If we could automate those three scores, we could surface candidates automatically and reserve human time for the final optimization

What we shipped

Keyword-landing audit pipeline

• Pulls Google Search Console + Ahrefs data nightly, clusters page-2/3 keywords by quick-win potential
• Intent classifier (Claude) tags queries: navigational, informational, transactional, local
• Daily "top 10 opportunities" email per client - no more manual audit mornings

AI citation monitor

• Polls ChatGPT, Gemini, Claude, and Perplexity with branded and unbranded prompts
• Flags when a client gets cited - and when a competitor gets cited for a query the client could own
• 5 AI Overview captures logged in Q1 alone

GEO optimizer

• Rewrites target pages for answer-engine consumption: factual statements, structured data, FAQ schema
• Human-in-the-loop: founder reviews the rewrite before it goes live
• Preserves traditional SEO signals - no canonical breaks, no orphaned pages

Reporting dashboard

• Client-facing: rank movement, traffic lift, AI Overview captures in one view
• Monthly reports generated in 2 minutes instead of 2 hours

Why this stack

Rails for the app because the founder wanted to ship quickly and the team has the expertise. Python for the data pipeline because pandas + pyahrefs is the shortest path to reliable SEO data. Claude for the rewrites because the tone-matching was noticeably better than GPT-4 on agency-voice content. Airtable integration because the founder still uses it for client state - we met him where he already worked rather than forcing a migration

Outcome

• +32% organic clicks lift for Class Composer (flagship case)
• 5 AI Overview captures for clients in Q1 - previously zero tracked
• Page-2 → page-1 in 3-4 months is now a repeatable outcome across the book
• Agency can take on 2× the client volume without adding headcount