AI Workflow/Content Management

Inspiration Production System

Inspiration can be generated systematically, teaching you to build an inspiration pool for content creation. Are you tired of thinking about "what to post today" every day? By accumulating three sources of inspiration, using the right tags, and posting stable posts with personal characteristics, you can turn it into a system that can run on its own.

Inspiration production system: Three sources of accumulated content, AI simulated audiences, and hot topics converge into an inspiration pool. Mika uses a net to fish out the shining light bulbs from the pool.
Three sources of inspiration merge into one inspiration pool

In two consecutive classes, the students all asked the same thing: I am the principal of a one-person studio and also a bell ringer, and I feel tired when I want to publish topics every day. This article breaks down my own "inspiration production system": three sources of inspiration, how to accumulate them in daily life, how to get them by placing tags, and two instructions for actual operation. The characteristics of hot-selling products are hard to come by, but stable posts with ordinary levels and personal characteristics can be produced by the system. Who is

suitable for?
  • One-person company, self-media, people who want to run a stable community but are stuck on "what to post today" every day
  • People who have a bunch of old articles, notes, and screenshots on hand, but can never find them when they need to use them.
  • Knowledge workers who want AI to help publish articles, but find that the writing does not taste like their own.
What can you take with you?
  • A set of "three sources of inspiration" model: past accumulated content, AI simulated audience, hot topics, how to combine them
  • Three tagging techniques that can be learned from the pool, plus a tagging command that can be copied directly
  • A two-sentence workflow from seeing the meme to posting, and two free skill packs

I think about "what to post today" every day, and I feel tired when I think about it.

In two consecutive classes, students asked the same thing.

One person said: I am the principal of a one-man studio and also a bell ringer. I feel tired when I want to post topics every day. Another person said: If you want to post systematically, you hope that AI can directly rewrite a week's worth of posts based on past blog posts, and in the format and tone you want.

My answer to this question is: Yes, and inspiration can be stably produced.

Let’s make our expectations clear first: the characteristics of popular products are still rare. But stable posting of ordinary-level posts with personal characteristics can be accomplished by establishing a system. I call this system "Inspiration Pool".

Questions to be solved in the inspiration poolLet the decision of "what to post today" change from relying on willpower every day to fishing from a well-maintained pool.

Three Sources of Inspiration Pools

My inspiration pool has three sources of inspiration. It's mediocre when used alone, but its power lies in combination.

A three-circle intersection diagram of three sources of inspiration: past accumulated content, AI simulated audience, and hot topics intersect. The central glowing place is "inspiration", and Mika looks up in the lower right corner.
Three sources meet, the center is inspiration
District 1 The content I accumulated in the past

The crisp articles, in-depth articles, daily work logs, content taught in class, and questions answered by students are all materials that you have sorted out and can be reused. Every article you have written in the past can be written again from a different angle and for a different audience. The value is: you never have to start from scratch.

District 2 AI simulated audience needs

Ask AI to play the role of my various audiences: the boss of a one-person company, the administrator who is new to AI, and the lecturer who wants to transform, and let them talk about their respective problems. This step helped me come up with "questions worth testing" and then compare them with real messages, conversations, and questions asked by customers to decide which ones to write first. AI simulations are outposts, not stand-ins for real audiences. I used what I made myself AI Consumer Verification Skill package allows AI to generate ten audiences at a time to comment on my content.

District 3 Hot Topics and Current Affairs

What everyone is discussing recently and what topics have traffic. This area is responsible for "being seen". I don’t look for topics one by one by myself. I will ask AI to send several agents to check separately (one to check trends, one to check discussion popularity, and one to check what opponents are writing). Each agent will only bring back information with proven sources, and then I will pick. The complete setting of this "multi-agent query data" will be discussed separately. I will publish a tutorial later, but it is good to know this direction first. But the topic is just the entrance, not the content itself. Just chasing hot topics will not build trust. The topic must be able to connect you to your major.

The power lies in the combination:Use hot topics to pick up the points that everyone pays attention to, and then connect them back to my professional solutions.The topic is responsible for attracting people, the situation is responsible for making readers feel "this is talking about me", and the solution is responsible for turning traffic into trust. A topic can be paired with different old articles, and one old article can cater to different audiences. This is the inspiration library, which can be continuously combined.

How to accumulate this pool

The inspiration pool is not built when needed, it is accumulated bit by bit. Two habits:

The first habit: when collecting, write one more sentence

In the past, when we saw good articles, we would like them and save them. I would like to add one more sentence now: why I think it is good, what impressed me about this article, and what kind of problems and needs did I find it for.

This sentence is the "instructions for use" of this material. What's the difference? If you only collect materials, you won’t be able to remember why you saved them in the first place three months later; if you write a sentence of materials, the AI ​​will know “what kind of data will be used in what situation” after reading it, and it can retrieve it when needed and use it in the right place.

The second habit: throw it away and let AI tag it

When you have an idea, see a good case, or a customer asks a question, you don’t need to sort it out. Just throw it to the AI and ask it to label it and return it to its place. When it’s time to post, I ask it to give me five inspirational words to write. I have accumulated a lot of small inspirations, which is an in-depth article, a live broadcast, and a sharing session.

The tag is the harvesting system for the inspiration pool

After the pool is raised, what really determines whether it is useful or not is the label. It is easy to throw the material in, but whether it can be fished out after three months depends on whether the label is well placed. Three tips:

Tip 1: Add at least three tags to one piece of material, one for each category

Align the three categories of Tip 3 below: one for theme, one for audience, and one for situation. You need to add one or two more. As soon as the tag is hung, the scattered materials will grow into a series of theme lines. When you want to write about a certain theme, pull out the whole series to become a material package.

Tip 2: Use fixed words for labels, don’t use synonyms randomly

This is the trap that most people fall into. Save "AI applications" today, "artificial intelligence applications" tomorrow, and "AI tools" the day after tomorrow. The same topic is split into three tags, and the pool is no longer profitable. The method is very simple: open a "My Tag List" and stipulate that each concept uses only one fixed word. Ask the AI ​​to read this list first when placing tags. If the words are not in the list, you will be asked first before adding them. This is the vernacular version of controlled vocabulary and my entire set of Tag link method Tag Wiki The starting point of .

Tip 3: Labels are divided into three categories to gain dimension when fishing.

My own habit is at least three categories: theme (what is this article talking about, such as "inspiration pool" and "skill package"), audience (who is this article written for, such as "one-person company" and "lecturer"), and situation (when will it be used, such as "opening story", "counterexample" and "data support"). If all three categories are available, you can issue a compound command: "Help me find the subject of the "Skill Pack", write it to the "One-Man Company", and use it as material for the "Opening Story"", and the pool will instantly become a database that can be queried.

In class, I gave an instruction that can be copied directly (paste your tag list together, and the AI will follow the fixed words):

This is my tag list (posted here). Below are three sentences about my brand and some scattered notes. Please give each article a topic, an audience, and a situation label. You can only use the words in the list. If there are no suitable candidates, let me confirm first. Let’s gather those on the same topic together and finally give me 5 words of inspiration to post this week.

Practical operation: from seeing the meme to sending it out

There is a premise before doing this step: first put past articles, logs, and student questions and answers into a place where AI can find it (a folder, a note library, or an AI project will do), then it will have something to harvest. The place is ready, all that's left is two sentences.

I saw a popular meme and thought I had some ideas, so I threw the meme to the AI:

This meme is very interesting and I saw inspiration in it. Please help me check which article I have written in the past is consistent with this, and help me write one.

It will pull out the right old articles from my database (this step is accurate only if you tag them well), and combine them with the memes to write a new article. I will say this after I finish writing:

Inspirational, but too professional. Please help me check the title from the audience's perspective and change the title into a version that my audience can understand.

Two sentences, one post. The memes are new, the content is what I have accumulated, and the titles are understandable to the audience. The three areas are in their place in this process: the third area is for memes, the first area is for content, and the second area is for title control.

The four-step process from topic to post: seeing hot topics, matching my past content, AI generating the first draft, AI consumer reviewing the title, and finally publishing the post
Actual operation flow: see the topic → match the old content → generate the first draft → AI review the title → publish the article
It gives you a draft, not a finished productThe two sentences omitted are "thinking about the theme from scratch" and "looking through old materials", but they do not omit your judgment. I will check the generated version again: whether the facts are correct, whether the tone is like me, and whether I have overstated my words. It’s your level that sends it out.

On days when I have no inspiration, I choose a topic like this

On days when I really have no inspiration, I don’t just think about it and just ask AI to help me search for hot topics. After searching it out, I asked it two questions: Which of these topics are the entry points that everyone is paying attention to? What can I do professionally? Please help me choose a topic. I choose a topic, connect it to my past content, and finally use the opening formula to put together several titles, score them from the audience’s perspective, and change them to the title I want.

Five-step topic selection process when I have no inspiration: I have no inspiration, ask AI to search for hot topics, AI helps me choose the topic (two columns: the entry point of public concern and my professional entry point), continue my past content, and change it to the title I want
The topic selection process when you have no inspiration: search for topics → AI selects topics from both public and professional perspectives → pick up old content → change the title
Two skill packs used in this guaranteed route (free download from the official website)

AI Consumer Verification: Let AI simulate your various audiences and evaluate your articles, titles, and product copywriting.
Change title title-rewriter: Change one content to multiple opening versions, and each title is accompanied by a green light, yellow light, and red light risk assessment, without sensationalism.

The combination of the two packages is equivalent to asking AI to generate ten audiences to ask questions, and then change the title according to their language.

→ Go to the skill pack download page

Frequency Philosophy: Continuity is more important than online beauty

Finally, let’s talk about the mental method. When it comes to social media posts, I think continuity is more important than online beauty.

I have also thought about making a video, but with my current time allocation, the effort of making a short video is enough for me to write five to ten short articles. The production cost of a short article is low, so I first choose to use short articles to maintain the rhythm. This is my choice, not that short videos are bad.

Instead of spending a month to shoot a video very seriously and then getting tired and not wanting to shoot the next month, it is better to have a stable frequency: I currently shoot one to two on weekdays and three to five on holidays. If you have posted on weekdays and have potential, post again on weekends with a different beginning.

No need to wait for inspirationWrite down "why it is good and what situation it is used in", put the right tags, and accumulate the three sources, and AI can help you more stably combine posts with personal characteristics. This frequency does not rely on willpower, but on the system.
AI WorkflowMarketing StrategyCrisp textSkill Pack DesignTag link method

Tools mentioned in this article

Every tool above was linked when it first appeared. Here it is organized into a list so that you can collect them all at once. All skill packages are free and licensed by MIT.

Classroom Source

This article is integrated from two lessons:

7/12 ChatGPT Work After-Class Briefing 7/8 AI Agent Editor’s Briefing

The three-zone model, labeling skills, and posting rhythm were all demonstrated in class.