Reading time: ~8 min
TL;DR
A repeatable AI writing workflow is built on saved prompts, role-based instructions, app-specific contexts, and a routine that doesn’t require you to think about how to use the tool โ only what to do with it. Page Jarvis turns your browser into a workflow-native environment where the same system works across Gmail, Google Docs, LinkedIn, forms, and everywhere else you write.
What you’ll learn:
- The difference between using AI and having a workflow
- How to build a prompt stack organized by task type and app context
- How to create role-based workflows for recurring writing tasks
- How to establish a low-friction daily routine around your AI writing system
Most people use AI for writing the same way they use a calculator: as a sporadic tool for specific moments when they need it. Open it, do the thing, close it, move on.
That’s not a workflow. That’s a transaction.
The people who get the most out of AI writing tools are the ones who’ve turned use into system โ where the tool fits so naturally into their browser environment that the question isn’t “should I use AI for this?” but “which part of my system handles this?”
This post is about building that system inside Page Jarvis.
What Makes a Workflow Repeatable
A repeatable workflow has three properties:
- Predictable inputs โ you know what kind of writing task is coming
- Consistent process โ the steps you take are the same every time
- Stored instructions โ your prompts and preferences are saved, not re-typed
Without all three, you’re still improvising every time. With all three, the system does the cognitive lifting so you can focus on the actual work.
Layer 1: Saved Prompts by Task Type
The foundation of any workflow system is a well-organized prompt library. From your saved prompts, build stacks organized by recurring task type.
Email Stack
Shorten emailโ reduce length while preserving key messageProfessional toneโ formal rewrite for external communicationsFriendly rewriteโ warmer tone for internal or casual outreachFollow-up draftโ follow-up version of an existing message
Content Stack
Simplify passageโ plain-language rewrite for accessibilityStrengthen hookโ punch up opening linesShorten paragraphโ tight editing for scannable contentExpand pointโ develop thin sections
Review and Edit Stack
Grammar checkโ polish and correctClarity passโ remove ambiguity and fillerTone matchโ adjust to match a specific voice or brand
When these prompts are saved and organized, executing a task becomes: open Page Jarvis โ select prompt โ done.
Layer 2: Role-Based Workflows
Different writing contexts call for different instruction sets. Build role-based workflows that encode the conventions and tone expectations for specific writing situations.
Role 1: Outreach and Cold Email
Input: rough idea or bullet points Step 1: Draft full email from notes Step 2: Shorten and make more direct Step 3: Strengthen subject line Result: send-ready cold outreach
Role 2: Content Marketing
Input: topic or headline Step 1: Outline key points for this topic Step 2: Expand point [selected text] (run per section) Step 3: Simplify dense passages Step 4: Shorten final draft Result: polished article or post
Role 3: Recruiter / Talent Outreach
Input: candidate profile or LinkedIn background Step 1: Draft personalized outreach message Step 2: Make it concise and professional Step 3: Add relevant context about the role Result: tailored recruiting message
Role 4: Support Response
Input: customer message or complaint Step 1: Acknowledge and summarize the issue Step 2: Draft a helpful, empathetic response Step 3: Shorten to essential points Result: professional support reply
Layer 3: App-Specific Contexts
The same prompt behaves slightly differently depending on where you’re writing. Use app-specific contexts to get better output from the same instruction.
Google Docs Context
Focus: long-form editing, document drafts, collaborative review Saved prompts: Simplify paragraph, Shorten section, Strengthen argument, Grammar polish Key behavior: edits stay in the Doc, iteration happens in context
Gmail Context
Focus: email composition, message revision, fast turnaround Saved prompts: Shorten email, Professional rewrite, Follow-up draft Key behavior: fast single-action rewrites, quick turnaround
LinkedIn Context
Focus: posts, comments, professional presence Saved prompts: Punch up hook, Shorten post, Make it engaging, Reply to comment Key behavior: platform-specific tone, engagement-optimized output
Notion Context
Focus: notes, documentation, knowledge capture Saved prompts: Summarize notes, Extract action items, Simplify this Key behavior: reading and comprehension assistance alongside writing
Layer 4: Building the Daily Routine
The best system is the one you actually use. Here’s how to build a daily routine that doesn’t require willpower:
Morning: Process Stack
- Open Gmail
- Run
Shorten emailon 3-5 emails that need trimming before sending - Use
Professional rewriteon any external-facing message - Result: inbox is cleaner and more professional before you start deep work
Mid-Morning: Content Work
- Open your Google Doc or article draft
- Run section-by-section refinements using saved prompts
- Use
Simplify passageon any dense paragraphs - Result: content is closer to publishable without a full rewrite pass
Afternoon: Outreach and Communication
- Open LinkedIn or Gmail for outreach
- Use role-based workflow for recruiting, sales, or partnership messages
- Run
Follow-up drafton any pending replies - Result: outbound communication is consistent and timely
End of Day: Review and Polish
- Review anything you wrote during the day
- Run
Grammar polishorClarity passas needed - Save any new prompts that emerged from the day’s work
- Result: everything you sent was reviewed and refined
What Turns a Tool Into a System
The transformation from tool to system happens when:
- You stop re-inventing instructions โ your prompts are saved
- Context is built in โ app-specific behavior is encoded
- Routine handles the logistics โ you think about content, not process
- The system remembers what you like โ tone, length, style preferences are saved
Page Jarvis is the infrastructure layer for this system because it works inside the browser where the writing already happens โ not in a separate tab that interrupts it.
Key Takeaways
- A repeatable workflow requires saved prompts, consistent process, and stored instructions
- Build prompt stacks organized by task type: email, content, review
- Create role-based workflows for recurring writing situations
- Use app-specific contexts (Docs, Gmail, LinkedIn) to get better output
- Establish a daily routine that handles routine tasks automatically
- The goal is a system that thinks for you so you can focus on the work
Next Steps
Try this: Pick one recurring writing task you do every day โ a type of email, a LinkedIn post, a doc update. Build a 3-step prompt sequence for it, save all three prompts, and run the sequence for a week. By Friday you’ll know whether the workflow is working.
Page Jarvis turns your browser into a workflow-native AI writing environment. Build your workflow system and stop starting from scratch every time.
Leave a Reply