From Two Hours to Three Minutes: How We Automated a National Radio Station's Show Preparation
A national radio station's weekly show preparation took two hours of mechanical work every week. We automated the entire process with a sixteen-node n8n…
By Laurent Maréchal · · 5 min read
Two hours. Every week. For a task nobody questions because it has always been done that way. Gathering the material, formatting it, structuring it according to a running order: it's essential, but it's mechanical. When a national radio station brought us this case, the answer was obvious. No need for sophisticated artificial intelligence, no need for a six-month project. A well-built workflow, and the problem disappears. Here is what we did, what got in the way, and most importantly what you can take from it for your own organisation.
The concrete problem: two hours of mechanical collection every week
Preparing a weekly show always follows the same logic. You have to gather material from multiple places, sort it, reformat it according to the running order structure, then produce a document that can be used directly in the studio or control room. This is not creative work. It is collection and assembly work. Repeated week after week, it nonetheless represents a significant volume of time: two hours per show, measured, not estimated.
The real cost is not just the time. It is also the mental load of having to remember, of not forgetting a source, of checking that the format is correct before sending the document. Mechanical work that demands human attention where it should not have to be.
What we built: a sixteen-node n8n workflow
The solution is based on process automation via n8n, an orchestration tool we use regularly for this type of case. The workflow has sixteen nodes and covers the entire preparation chain.
The trigger
The workflow launches automatically on a weekly schedule. Nobody has to think about it, nobody has to activate it manually. On the morning of preparation, execution starts on its own.
Multi-source collection
Several sources are queried in parallel. The workflow fetches the material wherever it lives, aggregates it and normalises it. This was the part that took the most time manually: navigating between sources, copying, pasting, checking consistency.
Formatting according to the show's format
Once the material is collected, the workflow applies the running order structure. The document delivered at the output respects exactly the expected format, ready to use, with no further editing required.
The error workflow
This is the detail that is often forgotten in this type of project, and that makes all the difference in real production. A global error workflow monitors the scheduled execution. If something fails, an alert goes out immediately. Nobody discovers at the last minute that the document is not there.
What got in the way: source normalisation
Building the workflow itself did not present any major difficulty. The most delicate part was normalising the sources. Each source has its own formats, its own structures, its own irregularities. Edge cases had to be anticipated: a source that does not respond, a format that changes, a missing piece of data. This robustness work is what distinguishes a workflow that holds up over time from a prototype that breaks at the first incident.
The error workflow is not a luxury. It is precisely because we anticipated failures that the system is reliable under real conditions, not just in a demonstration.
The result: three minutes, measured on the morning execution
Two hours reduced to three minutes. This figure is measured on the automatic morning execution, not estimated on a theoretical average. The preparation document is available without human intervention, in the expected format, at the scheduled time.
For the station's teams, this is not just time recovered. It is a mental load eliminated and reliability gained: the document will be there, correctly formatted, even if nobody thinks about it.
You can view this project and others in the same vein on our dedicated show preparation case study page.
How to transpose this principle in your organisation
This case concerns the media industry, but the principle applies anywhere a recurring meeting or event requires a preparatory brief. Executive committee, sales review, weekly production check-in: the logic is identical. The material already exists somewhere. Gathering it is mechanical. Automating it is feasible.
Ask yourself three simple questions. Which recurring meeting in your organisation requires a preparatory brief? Where does the material needed for that brief live? What is the expected output format? If you can answer those three questions, you have the ingredients for a workflow of the same type.
If your organisation operates in a sector subject to European AI regulation, be aware that automation solutions like this one may fall within the scope of the EU AI Act compliance audit offered by Audityo, particularly if they influence editorial or broadcasting decisions.
The difficulty is not technical. It lies in the precise mapping of the existing process and in the robustness of the implementation. This is exactly what we do during an AI diagnostic: identifying recurring mechanical processes, evaluating what is automatable, and building something that holds up over time.
What this project says about automation in general
AI is often associated with complex use cases: predictive analytics, content generation, anomaly detection. But a significant share of the value lies in much simpler cases: repetitive, predictable tasks that consume human time without deserving it. Automating a show preparation workflow does not make the headlines at tech conferences. It saves two hours a week, every week, without anyone having to think about it.
This is often where the return on investment is fastest and most measurable.
Conclusion
This project for a national radio station illustrates a principle we apply regularly: before looking for a complex solution, identify what is mechanical and make it automatic. Sixteen nodes, a schedule, multiple sources, an error workflow. Two hours become three minutes.
If your organisation has a recurring process that consumes time assembling material that already exists, it is very likely automatable. The best way to find out is to spend thirty minutes together talking it through. Request your AI diagnostic: the discovery call is complimentary, with no commitment, and you leave with a clear picture of what is achievable in your context.