A Content Engine That Takes Raw News to Published Article, Hands Free
From raw headlines to a live, illustrated, search-ready article, on a schedule, with no human steps per article. The pipeline pulls trending news from several sources, picks the strongest story, writes it in a controlled editorial voice, generates a matching image and publishes it, across multiple sites.
Before: a content treadmill that never stopped.
The operator, a digital publishing business, runs a portfolio of news and blog sites. Every one of them needs fresh, relevant, well-written content, every day.
By hand, that does not scale past a handful of sites. Writers spend the first hour of the day just finding what to cover. Two sites end up covering the same story. Quality slips the moment volume goes up, and freelance copy arrives in a dozen different voices. Miss a few days and a site starts to look abandoned to readers and to search engines alike.
The brief: a pipeline that takes a site from raw, real-time news to a finished, published, search-optimised article with no person in the loop, on a repeatable schedule, across many sites at once.
What we built: a newsroom compressed into a workflow.
On a schedule, the pipeline does what a small newsroom does: gathers what is happening, decides what is worth writing about, writes it, illustrates it and publishes it. The two hard problems were selection and quality.
Selection: the strongest story, not the first one
The pipeline pulls headlines from news feeds and a news data API in parallel, merges them into one pool and strips duplicates by normalising URLs. It then ranks the pool to surface the strongest current stories. A duplicate checker makes sure the same story is never covered twice, and pulls from an alternate source when a headline has already been used.
Quality: a voice with rules
Each selected story is classified by topic. That classification picks an editorial angle and a set of trusted sources to cite. A style layer injects the tone and the rules: dense, factual reporting, and a ban on the hollow connective phrases that make automated content obvious. The prompt instructs the model to write a long-form, well-structured article using the selected stories as source material and the trusted domains for citations.
Finish and publish
A matching header image is generated, and the finished article is pushed straight to the site's publishing platform through its content API.
One template, many sites
The whole workflow is a template. Adapting it to a new site means swapping the routing, the author identities, the target URLs and the credentials, so one proven pipeline is rolled out rather than rebuilt each time. On a related build for a financial vertical, the same shape runs continuously off aggregated feeds, publishing under consistent author bylines.
What changed: daily publishing at software cost.
Each site publishes fresh, relevant, search-ready articles on a daily schedule with no human steps per article, and the engine can be cloned to any site in the portfolio. Because selection ranks the strongest story and the writing stage enforces a defined voice, the output reads like reporting rather than filler.
Newsroom output, at software cost, repeated across many sites at once.
Day to day now.
Each morning the scheduled run fires. By the time anyone at the business checks, the day's articles are live: topical, cited, illustrated and on-voice. The team's job has moved from writing and scheduling to reviewing the style rules now and then, adding trusted sources and deciding which site gets the engine next.
The same pattern, in your business.
You probably do not need daily news articles. But most businesses have a content job that is repetitive, rule-bound and always behind: the blog that has not been updated since spring, the product descriptions nobody has time to write, the weekly update to clients.
- If you run a property or real estate business, the same pattern turns new listings and market data into area guides and listing pages in your house style, and the strongest of them into story angles for real estate PR.
- If you run an e-commerce store, it writes and publishes product descriptions from your catalogue data, consistently, as stock arrives.
- If you run a professional services firm, it drafts a weekly briefing on changes in your sector from trusted sources, ready for a partner to approve.
The engine is the easy part. Deciding what is worth automating, and what needs a human editor in the loop, is where an AI Consultation earns its keep.
AI Automation
This is AI Automation applied to publishing: a pipeline that turns raw input into finished, on-brand content on a schedule. We build the same engine around your topics, your voice and your sites.