Brand Kit glossary: stop AI captions mangling product names

Wrong product names in captions look amateur. Prospects notice. Internal teams lose trust in ‘AI captions’ and go back to manual subtitles.

Why the usual approach fails

• Re-edit every caption file by hand
• Live with wrong spellings because ‘AI is close enough’
• Maintain a separate glossary doc nobody opens
• Disable AI captions entirely

Manual fixes do not compound. A glossary that is not in the caption tool might as well not exist.

A better system

The winning workflow is simple: make or upload your video → put it on a branded Share Page (your video + a next-step button) → if you have a list, give each person their own tracked link with a Campaign → see who watched → only send the next video to people who engaged, using a multi-step video sequence.

• Brand Kit transcript glossary next to logos and colours
• Buzz words / preferred terms for speech that matters
• AI captions that lean on your dictionary
• Human transcript review for the last mile
• Same Brand Kit look across Studio jobs

How to run it in Vyrol

1. 1Open Brand Kit and add product names, people, and tricky terms to the glossary
2. 2Run AI captions on a real clip
3. 3Review the transcript — fix stragglers, add them to the glossary
4. 4Export and publish (Share Page when it is a client/sales send)
5. 5Reuse the Brand Kit so the next job starts smarter

What good looks like

• Glossary lives in Brand Kit — not a forgotten spreadsheet
• Caption accuracy improves as the dictionary grows
• Brand look and language stay consistent across jobs
• Review step still exists — AI assists, you approve

Who this is for

Teams with product jargon, founder names, or sector terms that generic speech models keep getting wrong.

AI captions only win when your brand language is in the system — not stuck in someone’s head.