Blog Publishing Workflow With Claude AI: Cutting Time by 36%

This workflow describes the process of transferring the final draft of the blog post from Google Drive (the writing platform) to WordPress (the publishing platform).

This process includes the following substeps: formatting, final edits, internal/external linking, and identifying key highlights.

It’s the first iteration of integrating AI to optimize workflow and reduce content production time from a previously manual workflow.

The Goals of this Workflow Revamp

The goal of this blog’s publish workflow refinement is three-fold:

1. Workflow Refinement (Primary): Turn a habit-tested but manual workflow into an AI-assisted but human-controlled workflow to reduce the blog publish workflow completion time from 3+ hours to less than an hour.
2. Hands-on Learning in AI Creative Operations: Learn how to use Claude AI (free plan) effectively, identify its strong suits/limitations, and test the synergy between Claude and built-in AIs within my existing blog management tools (Google Drive —> Gemini, Notion—> Notion AI, Elementor —> Elementor AI)
3. Mental Offloading for Business Growth: Make space to begin social media marketing initiatives to increase subscriber count.

Why Revamp the Blog Publish Workflow?

• Established Consistency of Habit —> The blog publishing workflow has been well established and laid out, primed for AI prompting and automation that requires detailed step-by-step instructions to ensure efficiency.
• Highly Repetitive —> Easy to be taught via Claude skills and replicated via AI prompting with easy-to-measure outcomes
• Non-Strategic —> Won’t replace long-term strategic direction and ownership
• Immediate Impact —> One of the biggest bottlenecks right now, and easy to validate the level of satisfaction.

Why Claude AI?

• Teachable Tool with Structured Organization: I can teach Claude skills by providing process instructions in plain language, written as AI prompts that are easily retrievable and editable, to ensure my responses are consistent and more predictable.
• Advocated by Thought Leaders: Ali Abdaal, one of the most influential self-help influencers, mentioned Claude as a key contributor to the upscaling of his business’s creative operations.
• Tool Scalability:
  • All-in-one work hub with native connectors to existing management tools I already use for seamless API integration with intuitive ways to scale as business operations grow.
  • Business-friendly features such as data visualization, document artifacts, and custom dashboards (future expansion into Claude Code and CoWork).
• Audience Interest & Searchability: Claude AI is a trendy topic with high amounts of search volume.

Step 1: Current Process Evaluation with Claude

Step 1 was conducted on Friday, May 15, 2026.

I’ve connected Claude to Notion and Google Drive for this case study.

[Method 1] The Persona Approach: Task-Oriented AI Prompt with Task-Level Context Scope

1A. Where does this approach come from?

From: Anthropic Getting started with Claude.ai

Anthropic knows best how to approach its own product. I refined my AI prompting approach by following Anthropic’s guidelines for crafting effective AI prompts.

1B. AI Prompt Used

AI Prompt Formula: [Setting the Stage – What is your Role & What are Your Objectives?] + [Define the Task] + [Specify the Rules] + [Add Context Files]

Setting the Stage: You’re the creative operations specialist for the Nivationist blog website, responsible for improving the workflow process that turns the blog drafts from Google Drive into scheduled-to-publish blog posts in WordPress.

Define the Task: First, identify the friction points in my current workflow. My current workflow is shown in Iteration 0 (Manual Process). Second, suggest process refinements to remove friction points. Third, create a consolidated instructions document that helps me implement your process refinements in step 2. A successful revamped workflow is defined in the “Specify the Rules” section below.

Specify the Rules:

The revamped workflow completion time needs to be under an hour.

Create the consolidated document in Google Drive.

Don’t create new or update a Notion page unless specified.

Add Context Files: <Already linked above>

I’ve given Claude access to my Google Drive and Notion, so Claude can read the referenced documents and directly add pages.

1C. Claude’s Response to AI Prompt (B)

Claude AI auto-generates a new Notion page that shows the

[Proposed] Iteration 2 Process — Claude Workflow Analysis & Implementation Guide

• Full process evaluation of my current blog publishing workflow and
• Instructions on how to approach the next iteration of the workflow.

See Claude’s Sequential Steps Taken After AI Prompt Initiation (B):

Controlled permission is required: Claude AI looked into the main Notion page (not referenced in the prompt), where the referenced Notion was nested. It also auto-detects related pages (again, not referenced in the prompt) that may be relevant to the task to get the full context.

[Method 2] The Collaborative Approach: Goal-Oriented AI Prompt with Broader Context Scope

2A. Where does this approach come from?

Claude is advertised as a “thinking partner” by Anthropic, so I should talk with Claude in an effort to collaborate and encourage Claude to ask me questions. The back-and-forth interaction will give me a better idea of the true extent to which it could do.

My AI prompting approach is more conversational and gives Claude a broader scope of context for how this workflow revamp is a part of a bigger initiative.

2B. AI Prompt Used

AI Prompt Formula: [What I plan to do?] + [List of goals] + [The process] + [How the results should be presented?] + [Future collaboration expectations]

What I plan to do: Collaborate with Claude to help me document my process of how I turn my current content workflows into AI-assisted processes into case studies/operations blueprints/guides that will help my audience replicate my process, as well as position me as a hireable candidate in marketing, operations, and strategy in today’s job market.

The goal:

To create a case study/operations blueprint that I can showcase on my website (refer to the website link: https://nivationist.com/) to make me hireable as an AI operations expert, brand/content strategist, marketing coordinator, and, eventually, a creative director (refer to the shared PDF file: “YouTube Creative Director Ali Abdaal” for reference).

1. Refer to the PDF (with LinkedIn link): “YOUTUBE CREATIVE DIRECTOR Resume” to better understand my current skillset and meet the gap practically, realistically, and sustainably.

Create a downloadable and easy-to-replicate guide on this entire process (refer to the “The process” section below in this prompt) that is helpful for my target audience (refer to the website link: https://nivationist.com/ to know who my target audience is).

The process:

Let Claude know my current workflows: Link to Notion. Claude will let me know if anything else is needed.

Claude will ask me which workflows I want to refine. I will answer back.

Claude will fill out, show iterative changes, challenges, accomplishments, and actions we’ve taken to refine the workflow as Claude and I chat throughout the entire workflow refinement process.

• The entire workflow refinement process includes building the workflow, testing it out, troubleshooting (if applicable), and deploying it for use.

I will ask Claude to generate the results for download after completing the workflow refinement (refer to the: “The results” section below)

The results:

A completed case study/operations blueprint that makes me hireable in today’s job market in marketing & creative operations (do your own research on what’s needed for this case study to help me position myself for the roles mentioned above or emerging roles I haven’t mentioned, but will hire me full-time.

A completed guide that is easy for me to upload to my WordPress.org/Elementor website and an easy-to-replicate guide for my audience to download.

Both results should:

The brand voice of my website (refer to the website link: https://nivationist.com/)

Future iterative results

Each iterative change for the workflow in the future, after being initiated by me, will generate new result docs without overriding the previous result docs, but future result docs after the first must include a section that includes a simple timeline that showcases the iterative changes made to this workflow.

2C. Claude’s Response to AI Prompt (B)

Claude auto-generated the

• [Public] Nivationist Link Library on Notion as an expandable, single-truth links hub and
• A Google Docs artifact on the step-by-step instructions on how to implement and use the workflow with tailored AI prompts. ****
• Shows the full process evaluation of my current workflow within the chat (not as a Notion page compared to Method 1) for the coherence of current/future communication

See the Type of Responses Claude Gave After AI Prompt Initiation (B):

Claude proactively asks for specific contexts, allows me to pick what I want to do with a given list of recommendations, and gives me multiple-choice responses where I can put in my own alternative option for further processing, refinement, and human-driven operational control.

D. What I Learned from Step 1

• The more context you give to Claude and how this context is linked to your overall goal of completing the task that you requested Claude to do, the more helpful its responses will be to you in your use case.

• When you collaborate with Claude as opposed to giving Claude a task, you’re given more flexibility and control over how Claude approaches your request. You also partake in refining their thought process before they take action.

• Anthropic’s definition of how to create a good prompt is still relevant. The reason it yields a less varied response depends more on the level and amount of context provided. Although the Collaborative Approach doesn’t follow Anthropic’s structure for AI prompting, all the elements are still present (just with a different naming convention and more conversationally structured).
  • It’s best to combine the level of context and communication style from the Collaborative Approach + easy-to-use AI prompt structure from the Persona Approach.

• Although it may be helpful for Claude to conduct its own context search within a connected tool you’ve given access to, it’s a privacy, security, and control issue.

[Action] Ask Claude to limit their search only to the files and links you explicitly referenced and shared in the prompt. Be aware of the information stored and be honest about your level of exposure before connecting your applications to any AI. Protect sensitive information behind a password to restrict access or transfer sensitive information to another place, if necessary.

Step 2: Current Process Evaluation with Claude

Step 2 was conducted on Friday, May 15, 2026.

Why is human review necessary?

AI responses have a tendency to hallucinate, which means they will produce responses that sound like accurate answers but are made up. That’s where I would have to review the Claude artifacts to ensure their accuracy and intuitiveness of use for my use case.

This step is great for testing the synergy between Claude and other built-in AIs within my existing blog management tools (i.e., Google Drive —> Gemini, Notion—> Notion AI, Elementor —> Elementor AI)

The Human Review Process

The human review workflow diagram below was made by

1. Audio transcribing the process (via Google Keep) of how I conduct the human review process, and then
2. Generated via ChatGPT.

Switch Google Tasks with Todoist: Instead of using Google Keep (the default task execution tool when using Gemini), it’s better to route it to Todoist (the task execution tool I’m currently using).

[Action]: Research the best method to transport the extracted “Actions” tasks straight into Todoist from Gemini.

Step 3: Incorporating the “Next Steps” of the Newsletter Publishing Revamp

This step requires the testing of the Newsletter Publishing workflow revamp, which was completed on Tuesday, May 26, 2026.

Methodology

I am incorporating Claude into two workflow revamps simultaneously: blog publishing (this case study) and newsletter idea curation + publishing case study.

Currently, my newsletter creation + publishing schedule and my blog publishing schedule are on an alternating weekly basis.

Since Step 2 of this blog publishing revamp was in closer alignment with the newsletter schedule (week of May 18th), the workflow testing of the newsletter idea curation + publishing was done first.

This arrangement resolves several operational concerns:

• Reduce Operational Redundancy and Conflicts: By aligning the revamp test with its respective publishing schedule, I was able to get two deliverables while doing the test. A newsletter that’s published on schedule, and this case study.
  • This approach ensures that the experimentation required for this case study will have the least impact/conflict on the deadlines for more established deliverables (blogs/newsletters).
  • There is no need to wait until the week of May 25th (the next blog publish schedule) to start the workflow revamp. Proxy-based information could be gathered via another similar workflow revamp to anticipate friction, apprehend results, and increase the probability of succeeding the “under an hour” goal when the next blog publish schedule rolls around.

• Enriched Case-Study and Process Improvement by Building from the Previous: Since the approach and level of Claude-generated revamps for both workflows are about the same, redundant results and the lack of new findings could be resolved by feeding in post-process recommendations in the newsletter revamp test to improve the overall workflow revamp process and audience interest in this case study.

What was incorporated into this workflow from the previous?

The following two objectives from the newsletter idea curation and publishing workflow revamp will be incorporated into the workflow refinement/testing process in this case study.

---

Source: [See where the following tasks were derived from to get the full context.]

• “Build a consolidated, single-truth work execution system by teaching Claude skills and organizing context files into different projects. Use ChatGPT as a fallback when the Claude free limit is reached. Test Claude’s ease of use before paying for the premium.”
• “Experiment with Task Mining to have AI document my current workflows and auto-detect bottlenecks for future iterations, and save time writing down the workflow steps that help update skills in Claude.”

Step 4: Real-Life Workflow Testing

Sub-Step 1: Task-Mining Tools Research & Preparation

This sub-step was conducted on Tuesday, May 26, 2026.

Defining the Scope of Task-Mining for This Case Study

The long-term goal of incorporating task mining into the overall content management system is to remove a major friction point: having to document each workflow by memory and relying on my sole ability to explain the workflow steps in detail as the single point of truth and context for Claude to refine/suggest future iterative refinements.

With task-mining, I’m able to share my workflow either as a Loom/screen recording or have a dedicated task-mining software that can monitor what I do passively, record every click, keystroke, application switch, and mouse movement in real-time, which then automatically generates process maps and identifies bottlenecks/solutions from raw data for my review.

Because the Nivationist’s core philosophy and content are based on iterative case studies, incorporating task mining is crucial for process efficiency and scaling.

---

The purpose of this research/preparation substep isn’t to incorporate a fully operable task-mining workflow.

Because it is a part of a long-term initiative,

The incorporation of AI applications into the workflows isn’t yet fully established (still in the testing phase).

Instead,

The immediate goal/purpose of this substep is to

Research the type of tools I can use for task mining (get a sense of the learning curve, free-trial duration, after-trial pricing, pros/cons),

Set up a minimal task-mining application and gather process data whilst doing this case study to accumulate data to experiment with in a consistent learning schedule, and

Ensure this substep doesn’t affect the blog publish schedule for this week.

How do I approach task mining?

3 Task Mining Methods

Microsoft Power Automate: I’ve set up a free account to get a preview of the application and learn from its tutorials on task mining within its 90-day trial. It has a task-mining feature built in, but the synergy/intuitiveness between the current tools I use (Notion, Google Drive, Google Keep, Notebook LM, etc) and the enterprise-level functionality of Power Automate needs to be further explored.

Activity Tracker + n8n: n8n offers better synergy with the content management tools I currently use, but it lacks a built-in task-mining feature. Instead, n8n needs to pull raw data from an activity tracker application to generate results akin to native task mining. I’ve downloaded the Activity Tracker app so it can monitor my processes in the background and upload the raw data into Power Automate or n8n for process experimentation and business synergy to help decide which task mining platform to go with after the free trial ends.

Screen recording of my previous workflow (not the Claude-suggested workflow): There was an issue opening the Activity Tracker app. It also takes time for the app to load the activity data, and most of the data is uncategorized. Thus, I screen record my blog publish workflow as a precaution if Activity Tracker fails, and as additional context to close data gaps prior to Activity Tracker troubleshooting to ensure data coherence.

Sub-Step 2: Testing Claude-Suggested Blog Publish Workflow

This sub-step was conducted on Wednesday, May 27, 2026.

Below are the summarized instructions and AI prompts within the Google Doc artifact that Claude generated. The instructions are composed of three parts:

• Part 1: Proofread the final draft of the blog post, identify key highlights of the blog post, search engine optimization, and external/internal links suggestions to include in the blog post.
• Part 2: WordPress post formatting reference sheet, featured image generation with ChatGPT, and pre-publish checklist.
• Part 3: After-publish checklist.

---

Want a refresher on how I got the instructions + AI prompts? —> Navigate to Step 1

Part 1 — Before You Open WordPress

Copy your Link Library snapshot

Open your Nivationist Link Library in Notion. Copy the full Internal Links table + any Outbound sections relevant to today’s post topic. You’ll paste this into every Claude prompt below.

Run the Proofreading + Highlights prompt

AI Prompt Formula: [Role + Brand Voice Context] + [Specific Task] + [Structural Constraints] + [Post Content]

ROLE: You are an editor for Nivationist, a human-centered career growth blog written by Ming. Brand voice: honest, warm, and encouraging — speaks directly to hesitant starters building careers from scratch. Never corporate or preachy.

TASK:

PROOFREAD: Fix spelling, grammar, and punctuation errors. Do not change the author’s voice, sentence rhythm, or word choices unless something is genuinely incorrect. Return the corrected full post.

HIGHLIGHTS: Identify 5–8 sentences or short phrases that are the most insight-rich or emotionally resonant — the kind a reader would screenshot or underline. Return as a numbered list under “SUGGESTED HIGHLIGHTS” after the corrected post. No more than one per paragraph. Works for both paragraph and non-paragraph post structures.

BLOG POST TITLE: [PASTE TITLE]

BLOG POST: [PASTE FULL POST TEXT HERE]

What to do with the output: Accept or adjust corrections in your Google Drive draft. Use the Suggested Highlights list as your reference when applying highlights in WordPress — you make the final call.

Run the SEO prompt

AI Prompt Formula: [Role + Brand Voice Context] + [AIOSEO Field Constraints] + [Post Content]

ROLE: You are an SEO assistant for Nivationist, a career growth blog targeting people in the early stages of building a career who feel “not enough.” Posts are long-form, conversational, and human-centered.

TASK: Generate the following SEO elements. Keep all copy in Nivationist brand voice — warm, direct, never clickbait-y. AIOSEO free plan constraint: use exactly 1 primary keyword and make it appear consistently in the SEO title, meta description, and slug.

SEO TITLE (under 60 characters, includes primary keyword)

META DESCRIPTION (under 160 characters, includes primary keyword, written as a genuine value promise)

PRIMARY KEYWORD (1 phrase only)

SLUG (lowercase, hyphens, under 6 words, includes primary keyword)

ALT TEXT for the featured image (plain language description of what the image conveys)

BLOG POST TITLE: [PASTE TITLE]

BLOG POST: [PASTE FULL POST TEXT HERE]

What to do with the output: Paste SEO Title and Meta Description into AIOSEO fields. Confirm slug before publishing — never change after publishing.

Run the Link Suggestion prompt

AI Prompt Formula: [Role] + [Task + Link Format] + [Link Library Snapshot] + [Post Content]

ROLE: You are a content strategist for Nivationist. Your job is to suggest internal and external links that add genuine value — not links for the sake of links.

TASK: Read the blog post below. Suggest:

INTERNAL LINKS: 2–4 places where linking to another Nivationist page helps the reader go deeper. Format: “Suggested anchor text” → [URL] — Reason: [one sentence]

OUTBOUND LINKS: 2–3 places where an external tool or resource would genuinely help. Actively search for the best external resources — do not limit suggestions to the library only. Always include Readwise/Reader (readwise.io/reader) where relevant to reading, note-taking, or research habits. Format: “Suggested anchor text” → [URL] — Reason: [one sentence]

Only suggest links that are a natural fit. Do not invent URLs.

LINK LIBRARY: [PASTE YOUR LINK LIBRARY SNAPSHOT HERE]

BLOG POST TITLE: [PASTE TITLE]

BLOG POST: [PASTE FULL POST TEXT HERE]

What to do with the output: Review each suggestion — accept, adjust, or skip. Add accepted links in your Google Drive draft before copying to WordPress. Add any new outbound sources to your Notion Link Library.

[In-Process Friction Points / Solutions] Part 1: Post Proofreading, Identify Key Highlights, SEO, and Suggest External/Internal Links

Previous Workflow Time: 140 minutes

Claude’s Target Time: 22 minutes

Current Workflow Time (Actual Time): 10 minutes + Human-Review (~60 minutes) = 70 minutes

---

In-Process Friction Points / Solutions

Information Consolidation: Relaying the “next steps” proposed in the newsletter workflow, I took the initiative to create a new Claude project, feed all of the AI prompts into Claude’s project instructions, upload a copy of the full post for context reference, and include the Nivationist link library Notion page as a reference link instead of pasting each prompt one at a time in Claude in the general chat.

Human Review to Reduce AI Hallucinations + Errors: Because of the screen-recording countermeasure for task-mining, the blog post has been published ahead of schedule using my previous workflow (no AI). I can then use Claude to compare the no-AI-assisted published blog post workflow vs. the Claude-assisted blog post workflow.

1. I asked Claude to explain the differences between the human-revised blog post (I gave Claude the URL of the published blog post) and the Claude-revised blog post.
2. Claude has trouble detecting the actual highlighted sentences (with background in #FFD37A) on the actual published blog post after matching Claude’s conclusions with my own review.
3. I asked Claude to generate an error report, which led to the following errors in Claude’s response to my request in step a.
   • Matched sections = the same highlights that both the published blog post and the Claude-assisted blog post selected. Not used sections: the highlights that Claude suggested but were not highlighted in the published post. Human-added section: the highlights that I highlighted in my published blog post, but not recommended by Claude.
   • Claude is bad at detecting highlights in my published blog post. I found that the highlights Claude thought were highlighted insights are within a paragraph distance away from the actual highlights. Claude may be having trouble detecting the background hex code color used to identify the published highlights.

Claude provided flawed reasoning for the errors made. The root cause given by Claude for why these errors exist is also incorrect, as the published blog post is included in the initial request for a comparison report in step a.

Claude is great at refining grammar without changing the original post drastically. The Claude-assisted blog post is similar to my published blog post, with only minor word changes and sentence restructuring.

Claude’s SEO Title suggestion is too generic but SEO-safe.

[ACTION]: Claude needs to increase the number of external links. It suggested only two external links for an 18-page blog post.

Conclusion of Step 4 (Part 1): Next Step Approach

[ACTION] Add a criterion to the external/internal linking AI prompt to increase the number of recommended external links.

[ACTION] Use Claude’s SEO suggestions as a starting point. I will revise it further to emphasize how my blog post compares with other blogs on a similar topic, and I will use ChatGPT to generate additional suggestions to avoid hitting the Claude free plan limit.

[ACTION] Use Claude to proofread the entire post and identify key highlights. Then, I do a final review of the suggestions before publishing the post.

Part 2 — In WordPress

This sub-step was conducted on Wednesday, May 27, 2026.

1. Format using your Reference Card

Use the formatting rules below — you no longer need a published post open in another tab.

Element	Rule
Heading hierarchy	H2 for main sections, H3 for sub-points
Highlight color	[ADD YOUR EXACT HEX CODE — pending screenshot of live post]
Highlighted text	Colored background highlight — max 1 per paragraph
Paragraph spacing	[ADD: e.g. “20px bottom margin”]
Featured image	Full width, top of post, above title
Internal links	Open in the same tab
External links	Open in new tab
CTA at the end of the post	Link to newsletter signup + 1 related post

First time using this: open a published post, fill in the [ADD] fields above, then never open a reference post again.

• Add links

Using accepted link suggestions from Step 4, add them to the WordPress editor.

• All internal links open in the same tab
• All external links open in a new tab
• Newsletter signup CTA is at the end of the post

• Upload featured image

Generate in ChatGPT using this prompt:

AI Prompt Formula: [Task] + [Image Description] + [Brand Colors] + [Art Style Reference] + [Dos/Don’ts]

Task: Create a featured image for a blog post titled “[POST TITLE]” for a career growth blog called Nivationist.

Brand style: [DESCRIBE YOUR EXISTING IMAGE STYLE — e.g. flat illustration, warm muted tones, no people, geometric shapes, consistent with previous posts in this series]

Brand colors: [ADD YOUR HEX CODES]

The image should feel: encouraging, grounded, and human — not corporate or generic stock photo.

Reference the style of the previous post image: [DESCRIBE OR PASTE PREVIOUS IMAGE DESCRIPTION]

Dos/Don’ts: Don’t show text in images.

Tip: Save your image style description after your first post — copy-paste it each time so the series stays visually consistent.

• Final pre-publish check
  • Title matches SEO title (or intentionally different for readability)
  • Meta description filled in AIOSEO
  • Slug confirmed
  • Featured image uploaded + alt text added
  • All highlights applied
  • All links checked (internal same tab, external new tab)
  • Newsletter CTA at the end
  • Category and tags set
  • Scheduled for Friday, 10:00 AM EST
  • Read the post one final time, top to bottom

[In-Process Friction Points / Solutions] Part 2: WordPress Post Formatting, Featured Image Generation, and Pre-Publish Checklist

Previous Workflow Time: 57 minutes

Claude’s Target Time: 45 minutes

Current Workflow Time (Actual Time): 52 minutes

---

There are no in-process friction points, as the workflow revamp suggested by Claude moves all references and, previously, a mental checklist into a consolidated location. The time improvement is small, but it frees up mental space and ensures future consistency/accuracy.

Conclusion of Step 4 (Part 2): Next Step Approach

[ACTION] Full Scale Automation: This workflow is highly repetitive, has an established structure, and requires minimal human thought, which makes this process great for full automation using APIs to have the post saved as a draft in WordPress before being scheduled for publication after simple human review.

Part 3 — After Publishing

This sub-step was conducted on Wednesday, May 27, 2026.

Update your Link Library

• Add the newly published post to the Notion Link Library internal links table
• Note which series it belongs to and its post number
• Add any new outbound resources used to the relevant topic section

[In-Process Friction Points / Solutions] Part 3: After Publishing Checklist

Previous Workflow Time: N/A

Claude’s Target Time: 5 minutes

Current Workflow Time (Actual Time): 5 minutes

---

There are no in-process friction points, as the workflow revamp suggested by Claude is to update newly made internal links into a consolidated location for future ease of processing.

Conclusion of Step 4 (Part 3): Next Step Approach

[ACTION] Full Scale Automation: This workflow is highly repetitive, has an established structure, and requires minimal human thought, which makes this process great for full automation using Zapier or Make to create a trigger/result process node or custom APIs to auto-update the links library.

Step 5: Calculating Results

	Previous Workflow Time	Claude’s Target Time	Current Workflow Time (Actual Time)
Part 1	140 minutes	22 minutes	70 minutes (10 minutes + human review 60 minutes)
Part 2	57 minutes	45 minutes	52 minutes
Part 3	N/A	5 minutes	5 minutes
Total	197 minutes	72 minutes	127 minutes

It takes 50% less time to complete Part 1 of the workflow after factoring in human review.

Due to setting up a Claude project, uploading all relevant context files, and consolidating all AI prompts into a single skill, the actual Part 1 workflow took 10 minutes, but due to the human-review time that’s needed to double-check the accuracy of the AI responses, an additional hour is added to the total time spent in Part 1.

Calculation Method (exclude human review): % Change Between Part 1 Previous Workflow Time and Part 1 Current Workflow Time (Actual): (140-10)/140 100 = 92.9%

Calculation Method (total): % Change Between Part 1 Previous Workflow Time and Part 1 Current Workflow Time (Actual): (140-70)/140 100 = 50%

It takes ~93% less time to complete Part 1 of the workflow (post editing, highlight identification, link suggestion, and SEO).

The current process reduced the gap to the ideal goal of 1 hour from 137 minutes to 67 minutes, eliminating ~51% of the original gap.

Calculation Method: Gap Difference Between the Goal time of an hour and Previous Workflow Time: (197-60=137) – Gap Difference Between Claude’s Target Time (ideal) and Current Workflow Time (Actual): (127-60=67) = 70 —> % Improvement Towards Ideal Goal (70/137) 100 = 51.1%

It takes ~36% less time to complete the blog publishing process.

Calculation Method: % Change Between Total Previous Workflow Time and Total Current Workflow Time (Actual): (197-127)/197 100 = 35.5%

The % improvement is non-applicable for parts 2-3, but is foundational for full automation.

Step 6: Next Steps – Using the 3 Nivationist Principles

The 3 Nivationist Principles are what turn this philosophy into what we can take action on and feel the tangible effects of our progress from the inside out, despite feeling “not enough.”

Because ultimately, the number of great solutions available to us is endless, and the person working behind the plan isn’t a stone-cold robot but a living human with real needs and deserves a plan that works with them in mind.

The essence of Nivationist is “taking care of the human in us.” Rather than separating the emotional personal growth from the outcomes we bring, we should use our human tendencies (current work habits, reflexes, etc.) revealed within the process of our work to create a well-rounded plan on what to do next.

Nivationist Principle #1: Create a Critical, Self-Loving Mindset

The results of this process revamp attempt do not meet the 1-hour time limit goal. But it does improve my previous workflow time by an overall of ~36% but close to a reserved estimate of 60-70% if the human-review process is better refined and full automation is integrated for parts 2-3.

Mindset Shifts:

• Shifts in Perfectionism: From following processes for the sake of completion to using what was learned from previous processes to improve the current approach.
• Reimagined Perfectionism: Using the desire to improve as an iterative process and publishing case studies to report the process, regardless of whether the ideal goal was met or not. Using the results to build a better approach for the next attempt.
• Mindset Snapshot: “The approach may not align with the original instructions, but it’s at times necessary to have messy efficiency rather than inefficient completeness.”

---

Next Steps:

• Continue to iterate with AI-assisted tools to meet the 1-hour workflow completion limit consistently.
• Continue to publish case studies to showcase results to the public, regardless of whether the ideal goal was met or not.

Nivationist Principle #2: Build Personalized Routines and Systems

Work Habits / Friction Revealed:

• I like to retain my voice and the entirety of my ideas (letting AI only change the grammar, wording, sentence structure, formatting, etc.)
• I prefer to review the AI-generated results before publishing.

---

Next Steps:

[ACTION] Build a consolidated, single-truth knowledge hub on Notion. Readwise is already connected with Notion. Still needs to connect Google Keep, ChatGPT chats, and Snipd.

[ACTION] Add more criteria to the AI prompts to better align with task expectations.

• The variety of external media suggested should also be specified (including videos, podcasts, articles, etc., rather than just a certain category of media).
• The number of suggested external links must be proportional to the length of the blog post.
• The suggestions for grammar, wording, sentence structure, etc. must be shown as highlighted suggestions for easier review to eliminate the time it takes to identify what was changed.

[ACTION] Automate the Google Doc draft to WordPress draft post (include formatting, brand color incorporation, custom elements, etc.), automatically update the links library, and send me a Todoist task to review the post before scheduling to publish.

[ACTION] Experiment with the Task Mining methods (Power Automate, n8n + Activity Tracker by using both video screen recording and activity data from the tracker.

[ACTION] Research tools to support my human review process and cut down workflow completion time.

Nivationist Principle #3: Design Productive, Inspiring Living Spaces

N/A: This principle isn’t relevant in this case study.