A Simple Publishing Workflow for a Part-Time Hobby Website Owner

A part-time website owner can easily feel under pressure to publish more content.

You may see recommendations to publish every day, several times per week or according to an ambitious monthly target. Missing that target can make you feel as though you are falling behind.

That is not how I approach Action With AI.

A straightforward post might be completed in one session. A more involved article may require research, screenshots, practical testing or information from another organisation. That could take several evenings or even weeks.

A publishing workflow should help you turn a useful question into an article you are prepared to put your name to. It should not make you feel that you have failed because one properly researched post took longer than expected.

Start with the content plan

The next article should normally come from the website structure you have already planned.

Before beginning, establish:

• Which hub the post supports.
• Who has the problem.
• What question they need answered.
• What the article should help them decide or do.
• Which existing or planned posts connect with it.

This prevents the website from becoming a collection of unrelated articles chosen according to whatever subject happens to seem interesting that day.

The detailed planning method is covered in How to Plan Your Hobby Website Structure Before Publishing Dozens of Posts and How to Organise Your First Ten Hobby Website Posts.

Once the next question has been selected, the article-development process can begin.

Use ChatGPT as an interviewer, not a one-click writer

I use ChatGPT extensively to help produce articles for Action With AI.

I do not normally sit down and type the article myself. However, I also do not give ChatGPT a title and publish whatever it generates.

I treat the process as an extended interview.

We begin with the basic idea for the post. I then explain anything I believe may be relevant:

• My personal experience.
• Real-world examples.
• Mistakes I have made.
• What I think drives the reader’s problem.
• What beginners may misunderstand.
• Practical constraints.
• My opinions.
• Areas where I am uncertain.
• Questions I think the article should answer.

Some of my initial thoughts may be incomplete, loosely expressed or based on assumptions. That is acceptable at this stage. The purpose of the discussion is to bring the raw material into the open.

ChatGPT responds with follow-up questions and helps organise the ideas.

It may:

• Ask for a specific example.
• Identify an unsupported assumption.
• Separate personal experience from a general factual claim.
• Point out an apparent contradiction.
• Suggest an important perspective I have missed.
• Ask whether I have tested something.
• Identify a fact that may have changed.
• Recommend contacting someone who can provide a reliable answer.

The conversation continues until we understand what the article needs to say and what further work may be required.

Although ChatGPT eventually produces the prose, the process remains human-led.

The direction comes from the content plan and the reader’s problem. The experiences, examples and practical judgements come from me. I decide whether the conclusions are reasonable and whether the finished article accurately reflects what I mean.

ChatGPT helps extract, question, organise and communicate that material. I remain responsible for deciding what should be published.

Google’s guidance does not prohibit using generative AI to assist with content. It says creators should focus on accuracy, quality and relevance, particularly when content is generated automatically. Google’s guidance on generative AI content

The important question is not simply whether AI was involved. It is whether the finished article provides something accurate and useful to its intended reader.

Investigate important unanswered questions

The discussion sometimes reveals that an important fact must be verified before the article can be drafted responsibly.

Further work might involve:

• Reading official documentation.
• Comparing current services or prices.
• Contacting a provider.
• Conducting a practical test.
• Taking photographs.
• Recording screenshots.
• Measuring a result.
• Asking someone with relevant experience.

If an unresolved fact would materially affect the recommendation, the draft should wait.

This is an important difference between using AI as an editorial partner and using it as a quick content generator. The goal is not to produce a plausible-sounding answer as quickly as possible. It is to determine whether the answer can be supported.

How a backup question changed the article

The article about backing up a WordPress website provides a good example.

I use Wealthy Affiliate for hosting. I knew it created backups because Site Support had previously restored my website after I accidentally deleted important files through FileZilla.

However, while developing the backup article, several questions emerged:

• Why were WordPress backup plugins reporting problems?
• Did Wealthy Affiliate back up only the files or also the database?
• How frequently were the backups taken?
• Could I obtain a complete independent copy?
• Was relying on the hosting provider enough?

I did not know all the answers with sufficient certainty, so I contacted Wealthy Affiliate Site Support.

They explained that their platform intentionally prevents backup plugins from writing large archive files to the server. They also confirmed that Wealthy Affiliate backs up the complete website—both its database and files—every night.

Support can restore the site when necessary, and a member can request a complete downloadable backup. They generated one for me as part of their answer.

They responded to that particular support request within four minutes. That was my experience of one enquiry rather than a promised response time for every ticket.

This investigation improved the article because it replaced assumptions with specific information.

It also improved my own backup arrangements. The publishing process did not merely document what I already knew. It encouraged me to investigate something important that I had previously left to the hosting provider.

A reader using another host may receive a different answer. Their backup plugin may work normally, backups may be included with their hosting plan or the service may cost extra.

The article therefore needed to explain which questions readers should ask rather than assume everyone had the same arrangement as me.

Prepare the content brief and evidence plan

Once the discussion and necessary investigations are sufficiently complete, we prepare a formal content brief.

The brief normally defines:

• The working title.
• The intended reader.
• The reader’s problem.
• The central argument.
• The proposed structure.
• Personal experience to include.
• Claims requiring verification.
• Evidence already available.
• Evidence still needed.
• Images and screenshots required.
• Internal-link opportunities.
• Connected subjects that should not be repeated.
• The intended conclusion.

The evidence plan is particularly important when an article involves practical work.

If you write the article first and only later realise that a photograph, measurement or screenshot was needed, the original opportunity to capture it may have passed.

The detailed approach is covered in How to Build an Evidence Portfolio Before Writing a Blog Post. At this stage, the important point is to identify the evidence requirements before drafting wherever possible.

Draft and review the article

ChatGPT produces the first full draft after the discussion, research and brief have established what the post needs to contain.

The first draft is not automatically the finished article.

I review it by asking:

• Does it answer the intended reader’s question?
• Does it accurately reflect the discussion?
• Are factual claims properly supported?
• Is personal experience presented as experience rather than universal fact?
• Has AI overstated any conclusion?
• Is important context missing?
• Does the article flow in a sensible order?
• Are sections repeating the same idea?
• Has it gone too deeply into a subject covered by another post?
• Would a real example communicate the point more effectively?

Unnecessary repetition is a frequent issue. A long discussion may establish the same conclusion in several ways, and all those versions can find their way into the first draft.

The article can then be rewritten into a cleaner version.

This should not become endless polishing. A post does not have to be perfect. It needs to be accurate, clear, useful and sufficiently complete for the question it answers.

Add the evidence and promotional images

The written draft is only one part of the finished post.

I manually insert relevant screenshots, photographs and diagrams because they need to appear beside the information they support.

My usual process is:

1. Select the relevant image.
2. Give it a descriptive filename.
3. Convert or upload it as a WebP file.
4. Add meaningful alt text.
5. Add a caption when visible context would help.
6. Insert it beside the relevant passage.
7. Check its size and appearance.

Alt text conveys the meaningful content or purpose of an image for someone who cannot see it. A caption is visible to everyone and can explain what a screenshot shows or why a detail matters.

Not every image needs a caption, and alt text should describe the useful content rather than becoming a place to repeat search phrases.

I also create a Pinterest-friendly vertical image and currently use it for two purposes:

• The WordPress featured image.
• The Pinterest pin image.

Using the same design provides consistency. Someone clicking the pin arrives at an article carrying the same visual message.

A tall Pinterest image can occupy considerable space as a website’s featured image, so its desktop and mobile presentation still needs checking. If necessary, the same design could later be exported in two proportions while retaining the same colours, image and headline.

For now, reusing one asset keeps the process simple.

Transfer, link and check the article in WordPress

Once the final version is ready, my WordPress process is straightforward:

1. Create a new post.
2. Copy the finished article from ChatGPT.
3. Paste it into the WordPress editor.
4. Add the evidence images.
5. Set myself as the author instead of “admin.”
6. Select the appropriate category.
7. Check the title and URL.
8. Confirm the internal links.
9. Read the complete article again.
10. Check its presentation.
11. Publish it.

Because the post comes from an existing content plan, links to connected articles are often identified in advance and included naturally within the final text.

Before publishing, I still confirm that:

• The target article exists.
• The correct URL has been used.
• The linked words describe the destination.
• No placeholder remains.
• The post links back to its hub once that hub exists.

Publishing the new article may also create opportunities to add links from older posts. Those updates should be recorded as part of the publication work.

Reading the article inside WordPress matters because the finished layout can reveal problems that were less obvious in a chat window or document.

Check the heading hierarchy, spacing, lists, links, image sizes, captions and mobile display before publishing.

The ChatGPT draft is not the finished product. The post is finished when it has been transferred, attributed, categorised, linked, visually checked and published correctly.

Publish or batch the Pinterest pin

The cleanest workflow would be to create and publish the Pinterest pin immediately after the WordPress article.

That would close the complete production loop.

In practice, I tend to create pins in small batches. Working on several similar tasks together can be more efficient than continually switching between research, WordPress and Pinterest.

The risk is forgetting which published articles still need pins.

A simple content tracker can use statuses such as:

• Planned.
• Discussion and research.
• Evidence required.
• Drafted.
• WordPress published.
• Pin required.
• Pin published.
• Complete.

Batching is not a problem if the outstanding work is recorded and completed reliably.

There is no correct publishing speed

As a rough personal estimate, a relatively straightforward post may sometimes take me a couple of hours. That is not a target other people need to match.

Completion time depends on:

• How well you know the subject.
• The complexity of the question.
• The research required.
• Whether practical testing is needed.
• Whether another person or organisation must respond.
• The number of images involved.
• Your familiarity with WordPress and AI.
• How much uninterrupted time you have.

You might complete one article in a single session. You might produce several straightforward posts during one Saturday. An evidence-heavy article could take several evening sessions or weeks.

All those patterns are valid.

Frequency describes how often a new post appears. Consistency means continuing to return to the production process.

A part-time website owner can work consistently even when publication intervals vary. One week may produce several articles. Another may produce none because one complex post is still being investigated.

An unrealistic target—such as twenty posts every week—can encourage rushed research, weak evidence and superficial articles. Missing the target may also create a false sense of failure.

This does not mean quantity is irrelevant. A useful website still needs a growing body of content.

The aim is sustainable production without sacrificing the research, evidence and clarity that make each post worth publishing.

Build a process you can return to

The workflow can be reduced to six main stages:

1. Select and discuss the question.
2. Research it and collect the necessary evidence.
3. Prepare the brief and evidence plan.
4. Draft, verify and edit the article.
5. Add the images, links and WordPress settings before publishing.
6. Publish or queue the Pinterest pin and record what remains outstanding.

A useful workflow does not promise that every article will be finished quickly.

It provides a dependable route from an initial question to a published piece of work you are prepared to put your name to.

For a part-time hobby website owner, consistency means continuing to follow that route—not racing someone else’s publishing calendar.