Video

Nicole Junkermann on Checking an AI Match Report


Software can now write a match report within seconds of the final whistle. Nicole Junkermann explains why that report still needs a human check, starting with the score and the key facts, and why getting it right comes before writing it well.

Transcript


An AI match report still needs a human check. Check the score and key facts. Clear writing matters, but getting it right comes first.

Three key points


  • An AI match report still needs a human check.
  • The score and the key facts come first.
  • Clear writing matters, but accuracy matters more.

Fast drafts, careful checks


Automated writing works best where the inputs are structured, which is one reason sport became an early home for it. Scores, times and statistics arrive as data, and a system can turn them into readable sentences almost instantly. Nicole Junkermann sees real value in that speed, but she is clear about the order of priorities. A report with the wrong score or a misattributed moment does more harm than a slower one that is right. The engineering lesson is general: build the check into the workflow, not as an afterthought. Compare the draft against the source data, flag anything that does not match and keep a person responsible for what is published.

Read the full note on the blog: Nicole Junkermann on Checking an AI Match Report: A Short Video.

More videos


Find all six clips, the tech notes and the reading list on the Nicole Junkermann homepage.

All six videos → Browse the tech notes →