A creator’s project records become useful illustrated articles with AI assistance, followed by website review and approval
AI-generated illustration of this case: original projects become useful articles and a website with AI assistance, followed by review and approval. It does not depict Google’s actual review interface or establish its internal reasons for approval.

There is a gap between writing about personal experience and running a website that can display ads. Articles need pages, images and navigation; broken links need fixing. For someone hoping to earn from content, the initial obstacle may be less a shortage of ideas than the work of bringing these tasks together.

UPMI Lab worked with GPT Astra to turn completed research, teaching and everyday projects into public content. According to its operating records, the new Lab had 21 public articles when the first AdSense application, submitted for the existing company domain, was approved 13 days later. The change illustrated here is a way of turning experience into a working site, rather than the discovery of a universal approval formula.

Approval tips do not finish a website

It is natural to ask what approval requires. But treating article counts and word counts as targets can push readers' needs aside. Google's eligibility guidance calls for original content and compliance with its policies; it does not offer a universal passing threshold based on article or word counts.

Its official guidance also addresses access, distinctive content, user experience and navigation. The task is neither writing alone nor server configuration alone. Courses and professional services can help with that work. Someone else's success, however, cannot guarantee the applicant's approval or income. The cost of learning needs to be weighed against what the learner can actually complete.

Start with work already done

The operator initially planned to build a backlog of blog posts and attract traffic before applying for AdSense. Expecting that preparation to take considerable time, the operator instead turned completed work into a new website. HomeKit integration, editable presentations, classroom tools and NAS administration supplied the material. After choosing what to publish and setting the direction, the operator used GPT Astra to create text and images explaining each problem, approach and outcome.

Writing led into site implementation. Rather than purchasing a new root domain, the operator created the new lab.upmi.re.kr subdomain beneath the existing company domain. Astra helped build Korean and English pages, navigation and supporting links, check information and privacy pages, mobile display and public access, and repair unavailable documents. The operator supplied goals and material and reviewed results; Astra carried the work from drafting through website implementation and checks.

AI's role was not to anticipate a reviewer's decision. It helped connect writing, images and web work that the user otherwise had to handle separately. That did not remove the need for the operator's experience or judgment. Checking published claims and correcting errors remained human responsibilities.

What the 21 articles and 13 days actually measure

The application was submitted on September 16, 2026, and approved on September 29. The operator reports no rejection or resubmission between those dates. The Lab had 21 public articles immediately before applying, counting Korean and English versions as the same article. Its six homepage features were included in that total.

The domain submitted for review was the existing company site, upmi.re.kr; the Lab was a new subsite beneath it. The 21 articles therefore do not represent all content on the reviewed domain. Likewise, 13 days measure application to approval, not the time needed to create the company website and all the original project experience.

Google's internal reasons for approval are unknown. This case cannot establish that AI caused approval or that others will be approved on the same schedule. The documented outcome is that existing material became a new content site and its first application succeeded.

From waiting for income to operating a publication

Approval permits advertising. Whether readers return and whether advertising income covers operating costs and time are separate questions. This is a first-approval case, not evidence of sustained earnings. After approval comes the work of finding what readers value and improving the content.

Even so, the choices available can change. Someone wishing to publish their experience has an option beyond waiting to master planning, writing and production or outsourcing everything: completing a small site with an agent. Hands-on experience can also make it clearer where teaching or specialist help would be useful.

The possibility illustrated is a wider route from personal experience to public resources and services. Its value lies in an expanded ability to build, check and revise, rather than a promise of one-click income. The next question in monetization must move from approval tips to giving readers a reason to return.

Further readingExplore the implementation and supporting material (Korean) →
Sources & context

Analysis adapted from UPMI Lab's published account and operating records, alongside official Google AdSense guidance. Dates and article counts come from the operator's records; the discussion of practical capability is editorial interpretation. UPMI Lab and New Epoch Journal are both operated by UPMI Co., Ltd. This publication also displays AdSense ads.

AI assisted with drafting and editing. This article is part of our preview edition. Editorial standards & corrections →