Your expertise, not just your keywords.
Keyword tools tell you what the market is asking. They cannot tell an AI engine that you were on the floor the week the process changed. Writegarden captures that in the one format a busy operator still has time for: two minutes of talking.
Record or type · transcription and insight extraction · proposed topics you approve · straight into the article queue.
Why first-hand experience is the scarce input
Every competitor in your category can buy the same keyword data and prompt the same models. The result is a market where the median article is indistinguishable from every other median article — and AI engines, which increasingly prefer sources they can attribute to something with real standing, have less and less reason to cite any of it. What no competitor has is what happened inside your business this week: the objection that keeps surfacing on calls, the supplier change that moved your lead times, the sentence a customer used that reframed the whole problem. That knowledge rarely gets written down, because writing it down costs an hour nobody has. Speaking it costs two minutes.
How a memo becomes an article
The pipeline runs unattended between the recording and the approval. Nothing reaches your CMS without a person saying yes.
Record from anywhere
Speak into the browser on a phone or a laptop, or type the memo instead. Audio is captured in chunks, so a dropped connection costs you the last few seconds rather than the whole memo.
Transcription and insight extraction
The audio is transcribed, then read for what is actually load-bearing: claims, numbers, objections, anecdotes, and the questions the memo answers without being asked.
Proposed topics
You get a set of proposed articles, each with the reasoning for why it is worth writing and which insight it rests on. Approve, reject with a note, or leave it for later.
Approved topics enter the queue
An approved topic is generated like any other article — brand voice, imagery, schema, internal links — with the memo's substance carried through into the draft rather than summarised away.
The Experience half of E-E-A-T
Memo-grounded articles are where the Experience signal actually comes from. Paired with a real author profile on the Authority pillar, the byline has something underneath it.
Rejection is an input
Turning a topic down with a reason feeds the next round of proposals, so the system converges on the articles you would have chosen anyway.
A cadence that survives a busy week
The failure mode of every content programme is the week where nobody has a spare hour, followed by the month that quietly becomes a gap in the calendar. Writegarden assumes those weeks are most weeks. Two minutes of talking on a Tuesday produces a queue the rest of the pipeline can carry, and the publishing cadence holds even when your attention does not.
What operators ask about memos.
Do I have to record audio?
No. Typing the memo into the same box works identically — the pipeline needs the substance, not the sound of it.
Does the article just read my memo back to me?
No. The memo supplies the claim, the anecdote and the angle; the pipeline still does the research, the structure, the schema and the brand-voice pass. The memo is the ingredient a model could not have invented on its own.
What happens to the recording?
It is transcribed and stored against your workspace, isolated at the database level like every other piece of company data, and kept in the EU. The processing detail is set out in the privacy policy.
What if the proposed topics are wrong?
Reject them with a note. The rejection and the reason both feed the next round, so the proposals move toward the topics you actually want rather than repeating the same misses.