Turning a presentation into a video with a speaker

A video made of slides and voice is one of the simplest learning formats. It suits explanations, introductions, product news, policies and lectures. It plays anywhere, on the intranet, in a learning platform, via a link or on a website, and viewers need no special software.

The limit is the lack of interaction. You can pause and rewind a video, but you cannot ask it anything. For content where understanding must be checked or proven, a course in a learning platform is usually the better form; the video complements it as material to rewatch.

Each route has pros and cons; the choice depends on effort, how often content changes and languages.

A visible speaker can help: people pay attention to faces, and a speaker on screen gives the impression of a personal talk. But it can also distract if it covers content, constantly draws attention or is too large. The art lies in placement.

A few basic layouts have proven themselves: a small speaker in a corner when the slide is full; a half-figure next to the content when the slide has room; a large speaker for title and transition slides; and no speaker at all when a graphic needs the whole space. The right layout depends on each individual slide.

AI avatars allow a visible speaker without filming. They are created from a photo or short video and speak lip-synchronously with a synthetic or recreated voice. Changes and additional languages can then be made without a new recording.

Two rules are important. First: a real person may only be recreated as an avatar or voice with their explicit, documented and revocable consent. Second: viewers should be able to tell that speaker and voice are AI-generated. Article 50 of the EU AI Act contains transparency obligations for such content.

Every learning video needs captions. They help people with hearing impairments and everyone who watches without sound. If the video is generated from narration, the captions practically already exist; they just need the right timing.

Export in a widely supported format such as MP4. Consider whether one continuous video or one video per module fits better: individual videos are easier to find and update, a single video is more convenient to watch straight through. When something changes, only the affected part should need regenerating.

In Universe an avatar presents the course; the figure can be a default figure, a custom figure created from a photo or “My avatar” with your own selfie and voice, always with consent, and every face speaks only with its own voice. The speaker’s placement is proposed per slide, with a reason, and approved by the author. The presentation runs automatically and is recorded as one video per module or one continuous video; presenter and narration are labelled as AI-generated.