Putting a person on screen

Hi, I'm Mel, and I should disclose something before we start. I am an AI-generated presenter. Every video does not need someone like me. A face belongs only where it helps this particular message. This video gives you five choices, then shows how a presenter like me gets made.

You can film yourself. You can make a digital twin of yourself. You can create a fictional presenter from a description. You can use a public presenter supplied by a platform. Or you can use no face at all. Each is a production choice for a different situation. Choose whichever helps the viewer understand and trust the message.

A demonstration, screen walkthrough or visual explainer often works better without a presenter. The screen is already doing the teaching. A face earns the space when the person carries authority, reassurance, personality or a relationship with the audience. If it adds none of those, use the room for the material.

In HeyGen, creating your own avatar begins with two different routes. A digital twin uses footage of a real, consenting person. A virtual character begins with an image or written description. You can also skip both and choose from the public presenter library. Start by deciding whether the identity must belong to a real person.

Maya began as a written description: a warm professional presenter, dark brown hair, a green blouse, natural make-up and a confident expression. HeyGen returned four interpretations. We chose one and saved her beside Mel and Mo. Being specific guides the result, but it does not design every feature. You still inspect the options and choose the identity you want to keep.

Next come voice and motion. Maya uses a public voice. The motion menu then offers three engines, and they are not three kinds of avatar. Avatar Three is the simpler, lower-cost lip-sync route. Avatar Four adds script-adaptive motion and suits fictional or stylised characters. Avatar Five is designed for stronger consistency and more natural behaviour, especially with a real person's digital twin. Preview a short line before committing a full script.

Once the presenter exists, the workflow is ordinary. Finish the script. Choose the avatar, voice and motion. Split the message into scenes. Add the supporting visuals. Preview short sections, generate a draft, watch it from beginning to end, then revise. The useful advantage is not that the presenter becomes indistinguishable from a person. It is that one changed sentence does not require another camera session.

Synthetic presenters are strongest when they deliver clear, controlled information. Watch the small things: eye line, emphasis, hand movement, reaction and the timing between expression and words. If emotion, spontaneity or personal accountability is the message, film the real person. A polished avatar should not displace the human reason for watching.

HeyGen is a practical self-serve route for a branded presenter and repeatable revisions. Synthesia also has self-serve plans and is worth considering for structured training, with broader team governance higher in its range. Record yourself when your presence carries the trust. Use no presenter when the demonstration can carry the lesson. Each route serves a different job.

A real-person digital twin needs that person's permission. HeyGen requires the person to record a live consent statement, including when someone else is building the avatar. A fictional presenter avoids borrowing another person's identity, but it can still be mistaken for a real person. Disclose the use of a synthetic presenter when that fact could affect trust, consent or interpretation. These library videos use HeyGen. I began as a written description, and I am rendered with Avatar Three.

Begin with the message. Decide whether the viewer needs a person, then choose the least complicated format that does the job. If a synthetic presenter earns the space, test one short scene before you build the whole video.