by /
Three clients, three archetypes, one framework. Nick Dossa runs Vegas Auto Gallery, an exotic-car dealership. Jeremy Barker is a CEO at Murphy Door and host of the 90 Proof Wisdom podcast. Ardmor Windows & Doors is a $5M home-services company. Each ran through the same four-phase BlitzMetrics Personal Branding process. This case study is the team-side companion to Dennis Yu’s framework reflection from the same call — a side-by-side look at what AI did, what humans did, and the patterns that broke.
The Personal Branding definitive article is the canonical reference for the framework itself, this article just shows what happens when you run that same four-phase process on three very different clients. Phase names below come straight from /personal-brand, not a new system.
| Phase | Nick Dossa Vegas Auto Gallery | Jeremy Barker Murphy Door / 90 Proof Wisdom | ARDMOR Windows & Doors |
|---|---|---|---|
| Phase 1: Digital Plumbing | Existing site, weak schema, name domain in place | Multiple sites, fragmented entity signals across Murphy Door + 90 Proof Wisdom | Hundreds of auto-generated city pages, broken NAP consistency |
| Phase 2: Content | YouTube interviews on building Vegas Auto Gallery, Pagani collection content, Las Vegas philanthropy coverage | 90 Proof Wisdom podcast back catalog, Murphy Door speaking clips, community appearances | Real project photos and homeowner stories from completed installs |
| Phase 3: Authority | Press mentions on philanthropy, exotic-car community endorsements | Podcast guest network from 90 Proof Wisdom, executive references from Murphy Door | Five real project case studies replacing generic city pages |
| Phase 4: Knowledge Panel | Claim in flight — follow-up post will document panel state | Claim in flight — follow-up post will document panel state | Tracking neighborhood-level local pack post-cleanup |
Case 1: Nick Dossa, Vegas Auto Gallery
What we found at intake
Nick had Vegas Auto Gallery’s site doing double duty as his personal authority signal, but the About page read like generic dealership copy and ignored the actual story of how he built the business. The inventory pull surfaced YouTube interviews about his entry into the exotic car market, Instagram content featuring his Pagani collection, and Las Vegas press coverage of his philanthropic work, none of which appeared on the existing site.
What AI did
The AI agent handled WordPress page scaffolding, menu setup, and placeholder structure for About, Media, Partners, and Knowledge Panel pages. It pulled rough draft text from the inventory but produced generic bridging language between the real quotes, text that had to be cut.
What humans did
The team replaced generic dealership copy with the actual origin story sourced from Nick’s own interviews, embedded specific Pagani interview clips, and added a Las Vegas philanthropy section with press citations. Every claim on the rebuilt About page traces back to a public source.
Case 2: Jeremy Barker, Murphy Door and 90 Proof Wisdom
What we found at intake
Jeremy split entity signals across two domains and a podcast. Murphy Door covered his executive role; 90 Proof Wisdom covered his thought leadership; neither site cross-linked to the other and Google had no clean signal for which entity was canonical for “Jeremy Barker.”
What AI did
The agent inventoried 90 Proof Wisdom episodes and tagged each guest, generated draft show-note pages, and produced sameAs schema linking the personal brand site to Murphy Door and 90 Proof Wisdom, the technical work that clarifies entity relationships for Google.
What humans did
The team picked which podcast episodes to feature, wrote the connective copy that explains how Murphy Door and 90 Proof Wisdom relate, and built the link-back loop so each podcast guest’s site links back to Jeremy’s personal brand hub. This is entity linking applied at scale across a podcast network.
What is measurable now
Each new podcast episode now produces a show-note page that links to the guest, the guest’s site, and Jeremy’s hub, every episode strengthens three entities at once. The Knowledge Panel claim is in flight; we will document the panel state in a follow-up post once the sameAs schema has been live long enough for Google to reconcile.
Case 3: ARDMOR Windows & Doors
What we found at intake
Ardmor is the outlier in this matrix, a $5M home-services company, not a personal brand. Their site had hundreds of auto-generated “city pages” boilerplate landing pages targeting every service area in their region with thin, duplicated content. Google was treating the site as a thin-content factory rather than a credible local business.
What AI did
The agent identified and bulk-flagged the hundreds of low-quality city pages, mapped which neighborhoods had real installed projects, and produced page scaffolds for the legitimate service areas with placeholders for project photos and homeowner quotes.
What humans did
The team removed the spammy city pages and designed a plan to systematically create content for and to replace the spammy pages, with new, real, service area pages.
The pattern that held across all three
Three different client archetypes, one repeatable split: AI handled the structure, humans handled the substance. In every case, the agent produced WordPress scaffolding, schema, and inventory tagging in minutes. In every case, the substance, which interview to feature, which neighborhood was real, which podcast guest mattered, required a human who had been on the call.
The other constant: every site needed an inventory pass before any writing happened. Photos, podcast episodes, press coverage, completed projects, the raw material existed for all three clients before we touched the site. The job was organizing what already existed, not creating new content. This is the principle behind the inventory step in the broader SEO Tree framework and the Content Factory.
Dennis described the same pattern on the call in his own words: the agent runs for roughly ten minutes at a stretch, then stops or makes a mistake, and a human has to check the work and give feedback before it can keep going.
The patterns that broke
Three failure modes showed up across the builds. They are the ones to watch for if you run this process yourself.
- The agent linked to the wrong domains. When inventory tagging was sloppy, the AI dropped links to unrelated sites that happened to share a name token. Every external link needed manual verification.
- Generic bridging copy crept back in. Even with real source quotes, the agent produced filler sentences between them that read like every other AI-written page. Those sentences had to be cut, not edited.
- Schema and sameAs got out of sync with the visible content. The agent updated structured data faster than humans updated the prose, leaving pages where the schema described one entity and the body described another. Both layers had to be reviewed together.
None of these failure modes invalidate the AI-assisted build approach. They argue for the Website QA Audit as a non-optional step between agent output and publishing.
How to apply this to your build
If you want to run the same four-phase Personal Branding process on your own brand or a client’s, there are three paths.
- Do it yourself. The Personal Branding definitive article documents the full four-phase framework, Digital Plumbing, Content Production, Authority Building, Knowledge Panel. The Personal Branding Course adds templates, checklists, and walkthroughs.
- Have us do it for you. The BlitzMetrics and Local Service Spotlight teams build personal brand sites end-to-end using the same process documented above, including inventory, AI-assisted scaffolding, human substance editing, and Knowledge Panel claim.
- Run the AI agent path. The companion skill file below turns this case study into a reusable workflow for Claude — install it and the agent walks the same four phases on your inventory.
For comparable single-client builds, see how we built sites for Justen Martin, Tanner Laycock, Gavan Thorpe, and Ibrahim Awad. For the framework-side reflection from the same Oct 11, 2025 call, see Dennis’s Building an Evergreen Personal Brand. For the SERP-ranking side of the same problem, see How to Rank on Anyone’s Name.
Companion skill file
This article has a companion Claude skill file that turns the four-phase process into a reusable, automated workflow. After installing the skill, Claude can execute each phase on your inventory, building drafts, running the audit checklist, and producing deliverables that match the BlitzMetrics standard.
This case study links up to the Personal Branding definitive article and follows the standard set in the Definitive Article Guide. It was produced as part of the BlitzMetrics meta-article loop and the broader SEO Tree architecture, with cross-references to 9 Triangles and MAA.