Our weekly GEO audit said ChatGPT cited three clients zero times. ChatGPT was citing their own domains as the sole source. Here is the evidence that moved us off an index and onto live model probes, what it found that a citation count cannot, and the correction we owe.
Building in Public: The Free Quick Audit Pipeline — From Order to 15-Page Audit to Discovery Call
We made Quick Audits free across every Spotlight vertical, then built an AI pipeline that runs each one at max effort and books qualified owners into a 15-minute call. Here is the entire system, published per our building-in-public principle.
The File That Wouldn’t Attach: How Boomerang Blocks AI Agents in Gmail (and the DOM Fix)
Our agent kept failing to attach files to Gmail drafts with a cryptic “different extension” error — the culprit was Boomerang holding Chrome’s debugger channel. The fix: attach at the DOM level with file_upload. No clicks, no screenshots, no base64 through the model.
What We’re Sharing From Our AI’s Rulebook (And What We’re Not)
Historical architecture snapshot — July 2026. This article preserves an earlier three-memory implementation. The current system uses model-neutral, typed records with one authority per record class; provider-native memory is a convenience cache, not the global operational source. See the current shared-memory architecture. Right after the last article published, Dennis asked a good question: some of what’s sitting in Claude’s memory …
Why Your Claude Memory File Has A Size Limit (And What To Do About It)
Claude-specific historical incident. The size-limit lesson remains useful for Claude native memory, but native memory is a disposable cache, not the cross-agent authority. Durable context belongs in user-owned typed records and reaches each runtime through scoped adapters and verified packs. See the current shared-memory architecture. Right after I finished the admin-authorship fix, I told Dennis something he did not ask …
How We Launched Athlete Spotlight as a Repeatable Spotlight Vertical
A meta article documenting the step-by-step launch of The Athlete Spotlight, including domain purchase, DNS setup, lighthouse positioning, WordPress build, athlete showcase, and the $30/month Keap package.
How a Claude Agent Built Three Keap Quiz Pass/Fail Automation Campaigns
A Claude agent built three Keap quiz pass/fail automation campaigns for Content Factory Academy courses in one session. Here is the full process, cost, and compliance scorecard.
How We Migrated Content Factory Academy Courses to Free Access
How BlitzMetrics migrated 89 Content Factory Academy courses to free access in LearnDash and Keap, triggered 567 new enrollments, and built toward a program-outcome business model.
How a Claude Agent Set Up Keap Quiz Tags for LearnDash Certification
How a Claude agent created six Keap quiz pass/fail tags for LearnDash course certification across three Content Factory Academy courses, with a scalable naming convention.
How We Used an AI Agent to Build a Personal Brand Growth System (And How You Can Copy It)
Published on BlitzMetrics You’re reading the output of something we’ve never done before: we let an AI agent audit, strategize, draft, and package an entire personal brand growth campaign in a single session. Not a chatbot answering questions. An agent that ran searches, pulled Ahrefs domain ratings, cross-referenced 8 different search queries, identified gaps, wrote 10 pieces of publish-ready content, …
How We Disabled Comments Across 53 Legacy WordPress Sites (April 2026)
April 2026 case study: how we disabled comments across 53 legacy WordPress sites with SSH, WP-CLI, backups, and database verification.
From Brain to Bot: How to Turn Your Knowledge Into AI Agents That Scale Your Business
Learn the Brain to Bot framework: consolidate your knowledge, turn it into SOPs, and deploy AI agents to scale your business operations while you sleep.
- Page 1 of 2
- 1
- 2