The Content Engine
writes our briefing.
Newsletters, PDFs, podcasts and videos provide the source material. The relevant topics are in the newsletter the next morning.
- Text
- Newsletters, emails and documents
- Audio + Video
- Podcasts, recordings and transcripts
- Voice Notes
- Spoken notes and quick ideas
Incoming
NewsletterEmail inbox
Product PDFDocument
Podcast + VideoTranscript
Voice NoteSpoken note
Editorial run activeRating and curation active
Curate and rateCreate the newsletterDistribute
Edited and ready07:00 · Daily briefingWhat matters todaySelected storiesMeaning for an ownerA source on every claim Deep diveOne topic, properly explainedWhat happenedWhy it mattersOriginal sources Weekly reviewThe most important signalsDevelopmentsConnectionsWhat matters next
How it works
How our daily newsletter is made.
The workflow runs from the first source to distribution in four steps.
- 01
Collect sources
Newsletters, trade media, documents, podcasts, videos and voice notes arrive in one place.
Text - audio - video - voice notes - 02
Curate and rate
Only sources we trust make the briefing. We follow them ourselves and consider the people behind them leaders in their field. We then rate relevance, novelty and reliability. Repetition is removed.
Trust - relevance - novelty - reliability - 03
Create the newsletter
Selected topics are condensed, explained and connected to their original sources.
Summary - context - sources - 04
Distribute
The issue is published on Substack the next morning and delivered to subscribers.
Newsletter - web - archive
Behind the scenes
How the Content Engine is built.
Several technical and editorial steps work together, from daily source monitoring to the finished issue.
View the latest briefing on SubstackYouTube is monitored and transcribed daily
Selected YouTube channels are crawled and monitored each day. Videos are converted into searchable text through our own transcription API.
Newsletters enter the Content Engine inbox
Newsletters we subscribe to arrive directly in a shared inbox as source material.
PostgreSQL and embeddings form the RAG system
The content is stored in a PostgreSQL database with embeddings. This lets the system draw on sources and connections purposefully in later processing.
Multi-stage processing curates what matters
The engine processes sources in several stages, rates them and condenses them into the topics that belong in the briefing.
The selection becomes the newsletter issue
The curated material is assembled into the issue and its individual sections. The briefing is then published on Substack.
More possibilities
Four more jobs for the same engine.
Especially useful for companies that need to track markets, customers and new requirements.
01
Competitor radar
Sources
Websites, product pages, job posts, patents and trademarks
Output
New signals explained for product, sales and leadership
02
Tender monitor
Sources
TED, procurement portals and key buyer websites
Output
Relevant tenders with deadlines and a short fit assessment
03
Regulatory briefing
Sources
EU institutions, federal ministries, authorities and associations
Output
New requirements, deadlines and concrete actions
04
Customer voice
Sources
Support, CRM, reviews, meeting notes and voice notes
Output
Recurring questions, complaints and product requests
Free AI Strategy Call · 30 minutes
Where does the same work happen every day at your company?
30 minutes, free. Bring a workflow that costs you time every week. You get a straight read on what agents can do there, what they cannot, and what comes out at the end.