Choosing a voice
You set the voice when creating a template. In the New Template dialog, you’ll see a Use Premium Voice AI toggle. This controls the voice quality used across all slides, FAQs, and Q&A responses in the template. Pick the option that fits your use case — premium voices sound more natural but take slightly longer to generate.Regenerating audio after edits
When you save a slide with a changed narration script, the audio and word timings regenerate automatically as part of the save. You never have to remember a separate step to keep audio in sync with the script — if the save succeeds, the stored audio matches what you just wrote. If audio generation fails, the save fails too, so a slide is never persisted with a script that doesn’t match its audio. Saves that only change other fields — title, image, or context — don’t trigger any new synthesis. If you want to preview the new narration before saving, use the Preview button next to the narration script. It generates audio for the current script on demand — and skips regeneration on subsequent plays until the script changes again — so you only wait when there’s something new to hear. When you then save the slide, the editor reuses that audio instead of synthesizing it again. If you’d rather have the AI rewrite the script entirely from the slide image, click Use AI to Generate Script — this rewrites the narration based on what’s in the image, using your template context as guidance.How narration reads numbers and dates
AI Presentations normalizes narration text before synthesis, so numbers, dates, currency amounts, and phone numbers are spoken naturally instead of read digit by digit. Write values the way you normally would — “$12,500” is spoken as “twelve thousand five hundred dollars” — without spelling them out in the script. Some industry acronyms, such as “HVAC”, get a pronunciation override so the voice says them the way people do in conversation rather than reading each letter. Write the acronym normally in your script. The override only affects the spoken audio; captions and the transcript still show the written form. If an acronym in your scripts is read letter by letter and shouldn’t be, contact Demand IQ support to request a pronunciation override rather than respelling the word in the script, since the misspelling would appear in captions. Voice and delivery improvements apply when audio is generated, not retroactively. Audio generated earlier keeps its original delivery until you regenerate it, so a deck where only some slides were regenerated can mix older and newer delivery. To bring a whole deck up to date, save an updated narration script for each slide (or re-save each FAQ answer) so its audio regenerates.FAQ and Q&A audio
Audio for FAQ answers is generated automatically when you save a FAQ. If you edit the answer later, it regenerates automatically. For the Q&A transition and resume prompts (in the FAQs tab under Q&A Experience Settings), audio is not automatic — after saving the prompt text, click Generate Audio to produce the spoken version.Checking audio status
Throughout the template editor you’ll see status indicators on audio content:- Audio Ready — the spoken audio is generated and ready
- Generating… — audio is currently being produced
- Failed — generation didn’t complete; try regenerating