Technical SEO Agency Edition

Claude & Claude Code
for Technical SEO

A complete four-week training programme for SEO professionals — from smarter prompting to building production-grade automation tools.

🎯 Prompt Engineering 📚 Shared Prompt Library ⚡ Claude Code 🔌 API Integrations 🤖 AI-Powered Tools 🏗️ Team Standardisation
8
Tutorials
4
Weeks
7
Tools Built
2
Tracks

01 — What This Programme Is About

Technical SEO has always been a discipline that rewards people who can move fast, spot patterns in large datasets, and communicate complex findings clearly to non-technical clients. For years, those skills were built slowly through experience, and the work itself — crawling sites, triaging findings, writing briefs, checking status codes, reviewing schema — was unavoidably time-consuming.

Claude and Claude Code change that equation significantly. Not by replacing the expertise of a skilled SEO professional, but by removing the friction between having an expert insight and acting on it. The audit finding that used to take 20 minutes to translate into a client-ready paragraph now takes 90 seconds. The batch status check that required a custom script or a spreadsheet workaround now gets built in a single conversation. The content brief that required half a day of manual SERP analysis now takes 10 minutes.

"The best SEOs will not be replaced by AI. They will be replaced by SEOs who use AI better than they do."

— A useful way to think about the next few years of this industry

This programme is built around one central idea: Claude works best when the person using it brings deep domain expertise and clear thinking about what they actually need. The prompts, tools, and workflows in these tutorials are designed for technical SEO professionals who already know their craft — and want to spend less time on the mechanical parts of it.

02 — What Claude Can Do for a Technical SEO

Claude is a large language model with unusually strong reasoning, writing, and code capabilities. For SEO work specifically, this translates into four broad categories of value:

🔍
Analysis & Pattern Recognition

Paste a Search Console export, a Screaming Frog findings list, or a set of competing page structures and Claude will identify patterns, flag anomalies, and prioritise findings faster than manual review. It spots cannibalisation signals, CTR outliers, crawl budget waste, and structural inconsistencies across large datasets.

✍️
Writing & Communication

Claude writes well — clearly, technically, and adaptably. It can translate a 404 error into a plain-English client explanation, turn a list of audit findings into a prioritised executive summary, or draft a full technical SEO brief from a client intake form. Output that would take 30–60 minutes to write manually takes 2 minutes.

💻
Code & Automation

Via Claude Code, Claude writes, runs, and debugs Python scripts directly on your machine. It can build tools that check HTTP status codes across hundreds of URLs, parse server logs for crawl budget waste, follow redirect chains, connect to APIs, and generate HTML reports — all described in plain English, no prior coding required.

🧠
Strategic Reasoning

Given sufficient context, Claude reasons through complex SEO problems: should we consolidate or differentiate these cannibalising pages? What's the right content format for this keyword's intent? Which technical fixes will have the most impact given this client's constraints? It won't replace your judgment, but it's a fast, well-informed thinking partner.

03 — Before & After: What Changes Day-to-Day

The clearest way to understand Claude's value is to look at specific tasks and how they change. These are not hypothetical — they reflect what teams report after consistent Claude use over several weeks.

Day-to-day workflow transformation
Without Claude
  • Manually review and prioritise Screaming Frog exports row by row
  • Write client-facing issue explanations from scratch each time
  • Copy schema into Google's Rich Results Test, fix errors manually
  • Triage GSC data in spreadsheets to find declining queries
  • Check URLs one by one in PageSpeed Insights
  • Write technical SEO briefs over half a day
  • Identify internal linking opportunities page by page
  • Parse log files with specialised log analysis tools or Excel
  • Re-explain client context at the start of every analysis task
With Claude
  • Paste the Issues tab and get a prioritised action plan in 45 seconds
  • Run Prompt #4 and have client-ready copy in under 2 minutes
  • Paste JSON-LD and receive validated, corrected schema instantly
  • Upload GSC CSV and get declining queries, CTR gaps, and quick wins
  • Batch-check 100+ URLs via API and get a colour-coded HTML report
  • Run one prompt from a filled intake form — brief in 90 seconds
  • Automated tool scans sitemap and surfaces opportunities in minutes
  • Python script parses multi-gigabyte logs and reports waste patterns
  • Client Project Instructions mean zero re-briefing — ever

04 — The Time Saving Is Real

The following estimates are conservative and based on common agency workflows. Individual results vary based on task complexity and how well prompts are crafted — which is exactly what this programme teaches.

Prioritising a Screaming Frog crawl export
45–90 min
5 min
with Claude
Writing a technical SEO strategy brief
3–5 hours
20 min
with Claude
Schema markup review & correction
30–60 min
3 min
with Claude
Batch PageSpeed check (50 URLs)
2–3 hours
8 min
automated tool
Cannibalisation analysis from GSC data
60–90 min
10 min
with Claude
AI content brief from SERP analysis
2–3 hours
12 min
automated tool

Compounding gains: The biggest time savings don't come from individual tasks — they come from the accumulation across a week. A team of four SEOs each saving 2–3 hours per week is 8–12 hours of recovered capacity every week. Over a year, that's meaningful. Over two years, it's transformative.

05 — Why Claude Specifically

Several AI tools exist for SEO work. This programme focuses on Claude for specific, practical reasons:

📏
Long context window

Claude can process large documents in a single conversation — full crawl exports, lengthy log summaries, multiple competing page structures at once. You rarely need to break inputs into chunks.

🔬
Strong technical reasoning

Claude handles structured data, schema validation, log parsing logic, API response structures, and redirect chain analysis without getting confused by technical detail.

✒️
Excellent writing quality

Client-facing output from Claude is consistently clean, professional, and appropriately calibrated to the audience — which matters greatly in agency work.

Claude Code integration

The same model that analyses your GSC data can also write, run, and debug the scripts that automate collecting it — in one consistent workflow, with one subscription.

📁
Projects for team sharing

Claude for Teams lets your whole agency share Projects with persistent context — one setup per client means no one on the team ever starts from zero.

🛡️
Honest about uncertainty

Claude flags when it's making assumptions or when it doesn't know something with confidence — which matters when its output informs real client decisions.

An important note on accuracy: Claude is highly capable but not infallible. It can be confidently wrong about specific facts, and it doesn't know about recent Google algorithm updates or changes made after its training cutoff. Every tutorial in this programme includes quality gate guidance — what to verify before Claude's output is used in client work. Your expertise remains essential.

06 — Course Overview

This programme runs over four weeks and contains eight tutorials split across two parallel tracks. You do not need to choose one track — both are designed to work alongside each other, covering different aspects of using Claude in agency SEO work.

Efficiency Track For all team members — covers prompting, analysis, and workflows
Wk 1
1

Getting the Most from Claude: Prompting Fundamentals

The RRCF framework, roles, format control, negative instructions, Claude Projects setup

⏱ 45–60 min All team
Wk 2
2

SEO Workflows & Prompt Library

6 production-ready prompts: title tags, Screaming Frog, schema, client comms, content gaps, GSC

⏱ 60–75 min All team
Wk 3
5

Deeper Analysis with Claude

Cannibalisation detection, page structure comparison, full strategy brief generation

⏱ 60–75 min All team
Wk 4
7

Team Standardisation & Advanced Workflows

Projects architecture, SOPs, quality gates, prompt library maintenance, Show & Tell session

⏱ 60–75 min All team
Build Track For analysts and technically curious team members — covers Claude Code and tool building
Wk 1–2
3

Claude Code Setup & Your First SEO Script

Installation, authentication, agentic coding sessions, building an HTTP status checker

⏱ 90 min Build track
Wk 2
4

Building an Internal Linking Opportunity Finder

Sitemap parsing, page fetching, keyword matching, HTML report generation

⏱ 90–120 min Build track
Wk 3
6

Building API-Powered SEO Tools

PageSpeed Insights batch checker, GSC data puller, secure credential management

⏱ 2–3 hours Build track
Wk 4
8

Advanced Tool Builds & Capstone Projects

Log file analyser, redirect chain auditor, AI-powered content brief generator using the Claude API

⏱ 3–4 hours Build track

07 — The Tools You'll Build

By the end of the Build Track, every participant will have a fully functional, reusable SEO tools repository containing seven production-grade scripts:

Tool What it replaces Built in
HTTP Status Checker Manual URL checking, spreadsheet formulas Tutorial 3
Internal Linking Finder Page-by-page manual review Tutorial 4
PageSpeed Batch Checker Opening PageSpeed Insights one URL at a time Tutorial 6
GSC Declining Queries Puller Manual GSC export comparison in spreadsheets Tutorial 6
Log File Analyser Specialised log analysis software or manual Excel work Tutorial 8
Redirect Chain Auditor Manual redirect following, Screaming Frog crawl analysis Tutorial 8
AI Content Brief Generator 2–3 hours of manual SERP analysis and brief writing Tutorial 8

08 — How to Use This Programme

👥For team leads
  • Distribute tutorials at the start of each week
  • Run the Show & Tell session at the end of Week 4 (Tutorial 7)
  • Set up the Claude Projects architecture before Week 1 begins
  • Assign a prompt library owner from day one
  • Use the SOPs in Tutorial 7 as the agency-wide standard from Week 5 onwards
🙋For individual contributors
  • Complete Efficiency Track tutorials in order — each builds on the last
  • Do the exercises in each tutorial with real client data where possible
  • Build Track tutorials are self-paced — allocate extra time for Tutorial 8
  • Share useful prompts with the team as you discover them
  • Don't aim for perfection on first use — prompting improves with practice
📊For account managers
  • Focus on Tutorials 1, 2, 5, and 7 — highest immediate ROI for your role
  • Prompt #4 (Crawl Error Explainer) will change your client comms workflow
  • Tutorial 5's brief generator is directly applicable to new client onboarding
  • You don't need the Build Track, but attending the Show & Tell is valuable
🔧For technical SEOs & analysts
  • Complete both tracks in parallel — they reinforce each other
  • Build tools on real client data from the start — abstract exercises miss the nuance
  • Tutorial 8 is the capstone — block proper time for it
  • After Week 4, your next focus should be automating the weekly GSC and PageSpeed runs

09 — What You'll Need

💬
Claude for Teams

A Claude for Teams subscription. All Efficiency Track tutorials and the Build Track work via this.

Essential
💻
Terminal access

Mac Terminal or Windows PowerShell. Used to install and run Claude Code. No prior command-line experience needed.

Build track
🔑
Google API key

Free, from Google Cloud Console. Required for the PageSpeed checker and GSC integration in Tutorial 6.

Build track
🤖
Anthropic API key

From console.anthropic.com. Required only for the AI content brief generator in Tutorial 8. Usage costs are minimal.

Tutorial 8 only
🔍
SERP API key

SerpAPI or ValueSERP — both offer free trial credits. Required only for the content brief generator in Tutorial 8.

Tutorial 8 only
📊
Real client data

GSC exports, Screaming Frog crawls, server logs. The exercises work with fictional data but produce far more insight with real data.

Recommended

Before you start: Set up your Claude Projects structure (Tutorial 7, Section 2) before beginning Week 1. Having a client Project ready with context means every tutorial exercise produces work you can actually use, rather than throwaway practice output. Five minutes of setup saves hours of re-briefing.