Anyone who has ever tried to report live on a football match knows how challenging it can be. The action races on, important moments come thick and fast, and while you are still typing feverishly, the next exciting scene is already unfolding. Often, the quality and detail of the coverage suffer—or you simply run out of time, especially if you have to write the ticker on your own.
The Wormatia live ticker had been much the same until now: a dedicated writer put events into words and sent them off. But we wanted more. We wanted to make it possible to deliver precise, informative, and perhaps even slightly entertaining updates during hectic phases, without giving up full control over the text.
The solution? Artificial intelligence as a clever writing assistant.
The new button: AI power directly in the live ticker
The first step in our evolution was a subtle but effective adjustment to the live ticker interface. Next to the familiar text input field, there is now a new button. One click is all it takes to unleash the apparent magic of AI.
What happens in the background? When this button is pressed, the frontend sends the live ticker's current data through an interface (REST API) to our n8n installation. This transmission is deliberately kept lean to minimize delays—after all, fractions of a second matter in a live ticker.
In simplified architectural terms, it looks like this:
The aim of this UI extension was to integrate the writing process as simply as possible. Users should not have to navigate cumbersome menus or operate separate applications. AI assistance is available exactly where it is needed: at the moment of writing. So we continue to use the existing text field and just add one simple button.
The heart of the AI integration: The n8n workflow
The actual intelligence behind the text creation is hidden in a workflow within our n8n automation platform. This workflow is designed to understand the context of the match and generate precise suggestions.
Here is how it works:
Data intake and contextual understanding: As soon as the data arrives from the live ticker frontend (comment + ticker data via webhook push), the n8n workflow has access to the entire match chronology so far. It is "reading along," so to speak, and knows the goals, cards, substitutions, and all other relevant events. This comprehensive context is crucial so that the AI does not generate isolated snippets of text, but delivers meaningful additions.
The AI assistant does its work: Armed with this context, the data is forwarded to a Large Language Model (LLM), our AI model. We deliberately chose a smaller, cost-efficient model, since the task—writing short, concise live ticker updates—does not require enormous computing resources. The cost per generated suggestion is in the cent range, making the integration economically negligible.
The optimized prompt: The LLM receives a specially developed prompt tailored to the needs of live ticker coverage. We continuously refine this prompt, as ongoing use keeps giving us new insights into which wording is particularly helpful and where there is room for improvement.
Squad information as a trump card: Another important building block of our prompt is knowledge of the current squad. If the AI receives a player's name, for example, or even just their shirt number, it can use that information to enrich the text. The bare information "Yellow card #7" can thus become the suggestion "Klaus Müller is shown a yellow card after a foul."
The suggestion lands in the editor: The text suggested by the AI is then sent back to the live ticker frontend and inserted directly into the user's input field. The user now has full control: they can accept the suggestion unchanged, edit it however they like, or even send it back to the AI to get further variations or refinements.
The future of live ticker coverage is intelligent
With this AI extension, we have taken an important step in the further development of our Wormatia live ticker. I am convinced that this intelligent assistance is a valuable aid for delivering high-quality, informative live ticker coverage even during fast-paced phases of a match—while always retaining editorial control. Of course, the journey is not over yet, and I look forward to making the AI assistant even better in the future through further feedback and new ideas.
There are still ideas for extensions. A way to manage the lineup would be useful, for instance. Or more event types could be added.
It will not get boring.
Alla Wormatia!