Live captioning

Captioning is usually bought as an accessibility line item and then turns out to serve a far bigger share of the audience than anyone budgeted for — at an international trade event, the non-native speakers reading along often outnumber the audience captioning was specified for. In practice, you choose between human captioners (more accurate, especially with accents and jargon, and more expensive) and AI captioning (cheap, instant, and steadily improving but still prone to mangling exactly the technical vocabulary your event exists to discuss). Output goes to a dedicated screen, a lower-third on the main display, or attendees' own devices via the event app. Commercially, captioning is increasingly a procurement requirement for corporate attendees and a stated expectation in accessibility policies, so it's shifting from nice-to-have to table stakes for winning certain audiences. The common mistake is switching on AI captions without feeding the engine a custom vocabulary — product names, company names, sector acronyms — which is precisely where raw models fail and where errors are most visible on a six-metre screen. One honest nuance: captions become an asset after the event too; corrected transcripts feed session recordings, content repurposing, and search. Budget the cleanup pass, because raw live captions are rarely publishable as-is.