Business Builders · SEO Operating Standard

One process.
Every site.

Claude Code thinks. The site executes. Search Atlas keeps score.

The short answer: Claude Code runs every audit and every fix, on every platform. Replit, WordPress, and custom sites are just where the changes land. Search Atlas measures and reports, and never touches the site.

01 · The split

Three layers, one job each

Tools break when two of them change the same thing. Each layer owns its lane and stays out of the others.

Claude Code most of it Search Atlas some Replit / the site a little

Proportions shown are a working estimate for discussion. Confirm them with the time-tracking sprint. [TK: source needed]

Layer 1 · the brain

Claude Code

Diagnoses and fixes. Identical on every platform, which is what makes the process repeatable.

  • Full technical audit
  • Meta, schema, sitemap, robots
  • Redirects and URL mapping
  • Search Console data via Windsor
  • Pre-launch check
  • Copy recommendations to approve

Not thisRank tracking or client-facing dashboards.

Layer 2 · the scoreboard

Search Atlas

Measures and proves. It is our reporting system, not our editing system.

  • Keyword rank tracking
  • Local directories and citations
  • AI citation tracking, 30-day
  • Client-facing report data

Not thisThe on-page injection layer stays off. It sits on top of the site where our code cannot see it, and it overwrites approved Google Business photos.

Layer 3 · the hands

Replit & the site

Where changes land. This is the only layer that differs by platform.

  • Applies the fixes Claude Code writes
  • Build and design work
  • Publish and deploy

Not thisWe do not rely on Replit's own SEO audit. It reads the code, not the live web, so it misses Search Console errors and old-URL damage.

02 · Who owns what

The assignment sheet

Print this. Every SEO task has exactly one owner.

TaskOwner
Technical auditClaude Code
Meta titles, descriptions, schemaClaude Code
Redirects and URL mappingClaude Code
Search Console errors and queriesClaude Code
Pre-launch checkClaude Code
On-page copy recommendationsClaude Code → client approves
Applying the fixesThe site Replit / WP / repo
Keyword rank trackingSearch Atlas
Local directories and citationsSearch Atlas
AI citation trackingSearch Atlas 30-day, never daily
Google Business Profile postsGoogle, natively

03 · Platforms

The audit never changes.
Only the hands do.

You run the same command in Claude Code no matter where the site lives. It detects the platform and adapts.

Replit site

/seo-audit --apply-via replit

Claude Code audits, then hands the fix list to the Replit agent. Review, then publish.

WordPress site

/seo-audit --apply-via wordpress

Claude Code connects to WordPress and edits directly. Same fix list, same order.

Custom-coded site

/seo-audit --apply-via repo

Claude Code edits the code and redeploys. Same fix list again.

Build these two skills once
A monthly seo-audit skill and a pre-launch skill. Every person on every client runs the same two. That is the whole consistency problem, solved.

04 · The three levels

How we run it.
What the client gets.

These three levels are not separate products. They are the SEO level that comes with each client tier, per THE-ENGINE: fundamentals at Base and Foundation, full at Growth, comprehensive at Scale. SEO never consumes a play slot. One person owns the client and runs the whole loop.

Fundamentals

Base + Foundation · Quarterly · 2 hrs / mo

How to run it

  1. Run seo-audit in Claude Code
  2. Apply technical fixes only, no copy
  3. Check listings in Search Atlas
  4. Schedule Google posts natively
  5. Send the quarterly report

Client result

Found, and correctYour site is technically clean and your business shows up right in local search. This holds the ground. It does not grow rankings.

Full

Growth · Monthly · 4 hrs / mo

How to run it, plus

  1. Same loop, monthly instead
  2. Pull Search Console for 404s and queries
  3. Send copy recommendations for approval
  4. Apply what the client approves
  5. Report ranks and traffic monthly

Client result

Ranking and risingYou start showing up for the terms you sell on. Expect movement on a tracked keyword list and organic traffic that trends up.

Comprehensive

Scale · Monthly · 8 hrs / mo

How to run it, plus

  1. Blog and content optimization pass
  2. Review AI citations, 30-day report
  3. YouTube titles and descriptions
  4. Competitor gap check
  5. Report includes AI answers and video

Client result

Owning the categoryPage one for your target terms, named in AI answers, and video showing up alongside. Search your category, and there you are.

05 · Content

Content is a play.
SEO is not.

We already have two content plays. The gap is not that content is missing. The gap is that SEO and the content plays do not talk to each other.

SEO optimizes content that exists. Plays create content that does not. SEO fundamentals cover on-page hygiene on every page and post we already have, in every tier, and never use a play slot. Writing something new is Content Creation or Content Multiplier, and those cost a play slot at Foundation and up.

The handoff, once a month

Step 1

Find demand

Pull keyword gaps, competitor gaps, and AI citation gaps from Search Atlas.

SEO owner

Step 2

Merge the brief

Add those gaps to the answer backlog that already comes from sales calls and reviews.

AM

Step 3

Draft

Write it in client voice with the blog-writer skill. Run bb-house-style before it moves.

CM

Step 4

Approve and publish

Client signs off. Publish with meta title, description, schema, and an internal link to the service page.

AM → PM

Step 5

Track and feed back

New URLs go into rank tracking. Next month's gaps come from what did and did not move.

SEO owner

Steps 2, 3, and 4 are already written into the Content Creation and Content Multiplier plays. Steps 1 and 5 are the new part, and they are what this playbook adds.

What SEO owes content

  • A monthly list of what people actually search, from Search Atlas
  • The competitor gaps worth answering
  • Which existing posts to update instead of writing new
  • Meta title, description, and schema on every piece that publishes
  • Internal links from the new piece to the money page

What content owes SEO

  • The published URL, the day it goes live
  • The target question, so we can track the right term
  • A heads up before any URL or headline changes
  • Video titles and descriptions for anything on YouTube
  • Source recordings, so AI answers have something to cite

What content actually happens, by tier

Plays are bought per quarter, so content only runs if the client spends a slot on it.

Base

$2,000 · 0 plays

No new content. We fix meta and schema on what already exists. Nothing gets written.

Foundation

$3,500 · 1 play / qtr

Content only if they spend their one slot on it. Otherwise same as Base.

Growth

$7,000 · 3 plays / qtr

Usually one content play. Four net-new pieces a month, or the 8 / 2 / 2 multiplier package.

Scale

$15,000+ · 5 plays / qtr

Often both content plays, plus video. This is the tier where AI answer citations are realistic.

06 · Standing rules

Four things we never break

Copy needs a yes

SEO audits want to rewrite page text, and that text is often the most valuable fix. Send it to the client first. Never change approved copy on your own.

Report results, not effort

Clients do not care how many boxes we checked. Show them the search, and show them where they land. Rankings and traffic, every time.

Off the roster, off the tools

Remove an offboarded client from Search Atlas the day they leave. They burn credits and keep an open door into their site.

Time is the real cost

Search Atlas runs about $49 a site (figure cited in the Aug 4, 2026 team meeting). Our hours are the expense. When a client eats more time than their tier allows, move the tier, not the budget.

Business Builders
August 2026 · Draft for the next marketing meeting. Tier hours and skill command names are proposed, not final. Confirm owners and hours as a team before this becomes standard.