Most people use Claude like a search engine — short, vague questions that produce generic answers. This tutorial will teach you a fundamentally different approach: treating Claude as a **knowledgeable colleague** who needs context, constraints, and a clear brief to do their best work. By the end, you'll be writing prompts that produce dramatically better output for your SEO work.

## Learning Objectives

- Understand why context transforms Claude's output quality
- Apply the RRCF prompting framework to any SEO task
- Use roles, constraints, and output formats effectively
- Set up Claude Projects for persistent client context
- Build your first 3 reusable SEO prompt templates

## 1 Why Most People Prompt Badly

When we use a search engine, we've been trained to write short keyword queries. Typing "title tag analysis" into Google works because there are millions of pages to rank. Claude is different — it's generating a response specifically for you, right now, and it only knows what you tell it.

The quality of Claude's output is directly proportional to the quality of your input. Here's the same request written two ways:

### ❌ Weak Prompt

Analyse this title tag: "Buy Cheap Running Shoes \| Free UK Delivery"

### ✅ Strong Prompt

You are a technical SEO specialist working on a mid-size UK e-commerce brand selling running shoes. The target keyword is "running shoes UK". Review this title tag: "Buy Cheap Running Shoes \| Free UK Delivery". Flag issues with keyword placement, emotional appeal, and SERP truncation. Respond with: (1) a brief verdict, (2) the specific issues, (3) two rewritten alternatives with reasoning.

The second prompt gives Claude a **role**, **context**, a **specific task**, and a clear **output format**. The response will be incomparably more useful.

## 2 The RRCF Framework

Use this four-part framework for any non-trivial Claude prompt. You don't need all four parts every time — but the more you include, the better the output.

### R

**Role**  
Tell Claude who it is. "You are a technical SEO specialist with 10 years experience..." gives Claude a persona to inhabit, which shapes its vocabulary, depth, and assumptions.

### R

**Request**  
State clearly what you want. Be specific. "Review this" is vague. "Identify canonicalisation errors and rank them by estimated crawl impact" is actionable.

### C

**Context**  
Provide relevant background. Who is the client? What industry? What's the goal? What should Claude know about the audience or constraints?

### F

**Format**  
Specify how you want the response structured. Table, bullet list, JSON, a short paragraph, a client-ready email — Claude will match your format if you ask.

Let's see the RRCF framework applied to a real SEO prompt, with each component highlighted:

### Role

You are a senior technical SEO specialist.

### Context

The client is a B2B SaaS company selling HR software to UK enterprises.

They have 200 indexed pages and are targeting C-suite decision makers.

### Request

Review the following 5 meta descriptions and identify which best satisfy search intent for "HR software for enterprises". Flag any that are too generic, too long, or miss a clear CTA.

### Format

Respond with a table: 
| Page | Score /10 | Issues | Rewritten Version |

**Pro tip:** You don't have to write all four components in order. Claude processes the whole prompt. Write naturally — just make sure all four elements are present somewhere in your message.

## 3 Giving Claude a Role

Roles are one of the highest-leverage things you can add to a prompt. When you tell Claude "you are a technical SEO specialist", it shifts its entire response posture — it uses industry terminology correctly, it assumes you're familiar with concepts like crawl budget and canonicalisation, and it makes more expert-level judgments.

### Roles for common SEO tasks

| Task | Suggested Role |
| --- | --- |
| Technical audits | "You are a senior technical SEO specialist with expertise in crawlability and indexation." |
| Content briefs | "You are an SEO content strategist who writes detailed briefs for writers." |
| Client reports | "You are a senior account manager at a technical SEO agency writing for non-technical marketing directors." |
| Schema markup | "You are a structured data specialist familiar with Schema.org and Google's rich result guidelines." |
| Competitor analysis | "You are a competitive intelligence analyst specialising in organic search strategy." |
| Link building | "You are a digital PR specialist focusing on editorial link acquisition." |

## 4 Controlling Output Format

One of the most underused prompting techniques is explicitly requesting a format. Claude defaults to flowing prose, but for SEO work you almost always want something more structured. Here are the formats most useful for our work:

| Format request | Best used for |
| --- | --- |
| "Respond as a markdown table with columns: X, Y, Z" | Comparing multiple pages, keywords, or issues |
| "Respond as a numbered list, prioritised by impact" | Audit recommendations, fix lists |
| "Respond in two sections: (1) Issues, (2) Recommendations" | Audit summaries, reviews |
| "Write this as a client-ready paragraph, no jargon" | Client comms, reports |
| "Respond only with valid JSON in this schema: {...}" | When passing output to another tool |
| "Keep your response under 200 words" | Slack messages, executive summaries |

## 5 Using Claude Projects for Persistent Context

Every new Claude conversation starts with a blank slate — Claude has no memory of previous chats. For agency work, this is a problem: you'd have to re-explain your client, your methodology, and your preferences every single time.

**Claude Projects** solve this by letting you set a persistent system prompt that applies to every conversation within that project. Think of it as a permanent briefing document that Claude always has open.

### How to set up a Project for a client

In Claude for Teams, go to the left sidebar and create a new Project. You'll be able to write a Project Instructions document. Here's a template to adapt:

📋 Project Instructions Template

```
You are a senior technical SEO specialist working at [Agency Name].

CLIENT CONTEXT
- Client: [Client Name]
- Industry: [Industry]
- Website: [Domain]
- CMS: [e.g. WordPress, Shopify]
- Primary goal: [e.g. increase organic traffic to product pages by 30%]
- Target audience: [who their customers are]
- Key competitors: [list 2-3]

OUR AGENCY APPROACH
- We prioritise technical fixes before content work
- We use Screaming Frog for crawls, Search Console for performance data
- We present findings with a clear priority: Critical / High / Medium / Low
- Client-facing content should be jargon-free and solution-focused

TONE & OUTPUT
- Internal docs: concise, technical, direct
- Client docs: professional, clear, solution-oriented
- Always flag assumptions you've made if context is incomplete
```

## 6 Positive & Negative Instructions

You can steer Claude's output by telling it both what to do and what _not_ to do. Negative instructions are especially powerful for avoiding common frustrations.

Without negative instructions

Summarise the key technical SEO issues from this crawl data.

With negative instructions

Summarise the key technical SEO issues from this crawl data. Do not include issues with fewer than 10 affected URLs. Do not explain what each issue is — assume I know. Do not add a conclusion paragraph.

## 7 Practice Exercises

✏️ Exercise 1 — Rewrite a weak prompt

#### Take this weak prompt and rewrite it using RRCF:

"What are some good internal linking tips?"

1. Add a role appropriate to the task
2. Add context: pick a real or fictional client and their situation
3. Make the request specific and actionable
4. Specify a format for the output
5. Paste both versions into Claude and compare the results

✏️ Exercise 2 — Build a reusable template

#### Create a prompt template for one of these common tasks:

1. Choose your task: (a) reviewing a set of title tags, (b) turning a Screaming Frog error into a client explanation, or (c) identifying crawl budget waste from a list of URLs
2. Write the prompt with placeholder brackets for the variable parts — e.g. `[paste title tags here]`
3. Test it with real data
4. Save it to a Claude Project or shared doc so the whole team can use it

✏️ Exercise 3 — Set up your first Project

#### Create a Claude Project for a real client:

1. Open Claude and create a new Project from the sidebar
2. Use the Project Instructions template from Section 5
3. Fill in all the placeholder fields for one of your clients
4. Open a conversation inside the Project and ask: "Based on the client context, what are the three most important technical SEO questions we should be investigating?"
5. Notice how Claude's response reflects the context you provided

## 8 Summary

🎯

Use RRCF

Role, Request, Context, Format — include all four for best results

💬

Control Format

Always specify how you want the output structured — tables, lists, JSON, paragraphs

📁

Use Projects

Set up one Project per client so Claude always has context without re-briefing

**Key takeaway:** The effort you put into a prompt is almost always returned in kind. Spending 2 extra minutes writing a well-structured prompt can save 20 minutes of editing a mediocre response.