Content pipeline for the website

The client

The same Hungarian contracting company, whose website runs a professional news section. It serves two purposes: to show prospects that the company follows its field, and to keep the site regularly updated. The system was not built as an external engagement — it was built from the inside, as the company's process manager. The company name is not published.

The starting point

• Most industry news appears in international sources, typically in English. Going through them would have been a daily routine nobody actually had time for.
• As a result the news section updated in bursts: nothing for weeks, then several articles at once. A neglected section is worse than none at all — the dates give it away.
• Writing started from zero every time: find a source, read it, decide whether it is interesting, translate it, write it up, find an image for it.
• Keeping the house tone was a separate effort. An article thrown together quickly stands out immediately.

The solution

A daily routine that prepares the content but does not publish it. The decision stays with a person throughout.

1. Daily collection from 14 sources — international trade press, European and domestic sources alike. The system remembers what it has already processed, so the same story does not come back the next day.
2. Hungarian summaries — every item gets a Hungarian digest, so reviewing English-language sources does not require reading the originals.
3. The editor gets one email — with the day's list. One click opens a simple interface where they tick which items should become articles. That is the entire human effort: scan a list and tick a few rows.
4. Writing in the company's own voice — the writer does not work in generalities but follows a tone derived from the company's own earlier articles, taking search visibility and readability into account as well.
5. An illustration is produced too — every article gets a generated lead image, so the section does not read as a wall of text.
6. The article lands on the website as a draft — in the right section, but not published. Whoever approves it reads it, edits it if needed, and presses the button themselves.

Outcome

Writing no longer starts from zero. Someone who used to face a blank page now edits a finished draft — a different exercise, needing far less momentum.

The section updates regularly. Not in campaign-like bursts but continuously, because the obstacle was never the writing but the getting started.

The tone stays consistent. Articles do not stand out from the rest, because the writer works from the company's own earlier material.

🔑 And most importantly: nothing appears without approval. The system produces drafts, it does not publish. That is not a limitation but the point — only what a person has read and stood behind goes out under the company's name.

Where else this works

This component can also be deployed as a website feature rather than a standalone system, typically as part of building or rebuilding a site. The pattern works anywhere that:

• The company runs a professional blog or news section, but updates keep slipping behind other work.
• The field's main sources are in another language, which is why daily monitoring never happens.
• There is already an established voice — earlier articles the writer can learn from. Without that, the output will be generic.

One constraint matters, though: somebody has to approve. This system does not remove the need to deal with content; it makes dealing with it cost five minutes instead of two hours. Where nobody will review the output, this is not the thing to introduce.

Operation

The system runs on a daily schedule on a dedicated server. The source list can be extended, and the sample defining the tone can be refreshed from the company's own material if the direction of its communication changes.

This case study is based on a real system, anonymised. No figures are claimed for readership or search performance.