Content pipelines · Automate

Pages that gather and publish themselves.

Listings, articles and newsletters assembled from a source you already keep, on a schedule, with your approval in the middle. Built once, running after that without anybody opening a laptop.

What one looks like

Four shapes this usually takes.

They share one property: the facts already exist somewhere, and a person is currently moving them by hand.

Listings from a source you already keep

A spreadsheet, a database, a supplier feed or a public dataset becomes a page per row, each one a real page with its own URL, able to rank, rather than one page with a filter on it.

A newsletter that writes its own first draft

It gathers what changed since the last one, assembles it in your format, and waits for you to read it before it sends. Approval stays with a person.

Articles on a cadence you set

Drafted from your own material and your own positions, queued for review. You approve or you bin it, and nothing publishes without that step.

The thing you currently copy and paste

Pulling numbers out of one system every Monday to put into another. That is a pipeline with a human standing in the middle of it.

How we build it

One source of truth, one schedule, one approve button.

  • We find the source of truth, the sheet, the system or the feed that already holds the facts.
  • The pipeline reads it on a schedule and builds the pages, posts or emails from it.
  • Everything lands as a draft with your name on the approve button, until you tell us otherwise.
  • It runs on your own infrastructure and your own accounts.
  • When the source changes, the output changes. Nobody re-types anything.
  • You get the code, so a pipeline we built is one you can still run if we vanish.

The limits

What we will not build.

  • Volume is not the goal and it is not a ranking strategy. Google's spam policy names scaled content abuse specifically, and a thousand near-identical pages built from a keyword list is exactly that. We build a page per row only where each row genuinely differs.
  • A pipeline needs a source. If the facts live in somebody's head rather than in a system, the first piece of work is writing them down, and that is not automation.
  • Approval stays with a person by default. We will wire a fully automatic publish if you ask for it, after you have watched it produce drafts for a few weeks.
  • This does not make thin content rank. It removes the typing from content that was worth publishing anyway.

The one we run ourselves

ottown publishes on a schedule, and nobody opens a laptop.

A local news and listings site for the Ottawa area, ours, built on the pipeline described above. We carry the cost of it being wrong, which is the only version of this claim worth anything.

Content pipeline FAQ

Is this AI writing my content?

A model does some of the drafting on some pipelines, and on others nothing is generated at all. A listings pipeline is assembly, not writing. Either way the facts come from your source and a person approves before anything goes out. Which of those two a given job is, we tell you at the quote.

Will Google penalise automatically published pages?

Google's position is about quality and intent rather than about how a page was made. Pages assembled from a real dataset where each one answers a different question are ordinary pages. Hundreds of templated pages differing by a city name are the thing that gets hit, and we will not build those.

What if the source data is messy?

Usually it is, and that is most of the first week. A pipeline makes the mess visible all at once, which is uncomfortable and useful. A field that is blank on four hundred rows was blank before, nobody had looked.

Does it keep working if I stop paying you?

Yes. It runs in your accounts on your infrastructure and you have the code. There is a monthly if you want us watching it and fixing it when a source changes shape, and that is a separate decision from owning it.

Can I see one running?

ottown is ours. A local news and listings site for the Ottawa area that publishes on a schedule without anyone opening a laptop. It is the honest demonstration, because we carry the cost of it being wrong.

What does it cost?

A fixed one-off to build, quoted after we have seen the source and know how dirty it is, and an optional monthly to keep it running. You get a written price before anything starts.

02 Automate, the rest of it

  • AI chatbotsAnswering on your own content, not a generic model guessing.
  • Content pipelinesyou are hereArticles, listings and newsletters that gather and publish themselves.
  • Internal toolsThe spreadsheet-and-copy-paste job, turned into software.

What do you retype every week?

That answer is usually the whole brief. Tell us where the facts live now and where they have to end up, and we will tell you whether it is a pipeline or whether it is three hours of tidying a spreadsheet.

No deposit, and we reply within one business day.