✦ Summary
Avatars suit announcements and personal messages. For procedures and concepts, showing the thing usually teaches better than showing a face. Here is how to decide.
Do you need an AI video generator without avatars?
For most training video, yes. An AI video generator without avatars puts the process, the diagram or the product on screen instead of a presenter, and for procedures and concepts that is usually what the learner needs to see. Avatars earn their place in announcements, welcomes and personal messages, where a face carries trust.
This post is for L&D and training teams deciding which kind of tool to buy, or which kind of video to make next. It covers what avatars do well, where they get in the way, what replaces the face, and a decision table you can use per video.
What avatars do well
Avatar tools are good at a specific job, and it is worth being clear about it.
Announcements. A leader's update to a whole company reads as more personal with someone speaking to camera, even a synthetic one.
Personal messages at scale. A welcome video that greets each new starter, or a sales message recorded once and repeated for many accounts, is what avatar platforms were built for.
Soft skills role-play. Conversations about feedback, customer complaints or difficult calls need people on screen. A learner has to see a face react to judge tone.
Familiar format. Some audiences expect a presenter because that is what training has always looked like. Meeting that expectation has value.
If most of your videos are one of these, an avatar tool is the right choice, and you should pick one of the platforms built for it.
Where a presenter gets in the way of training
Most training teaches a procedure, a rule or a concept rather than delivering a message from a person. In those videos, a talking head competes with the content for the same screen and the same attention.
Attention goes to the face. People look at faces first. In a lockout procedure, the learner needs to watch the lock, the tag and the order of steps, not a presenter standing beside a small diagram. Richard Mayer's research on multimedia learning found that adding the speaker's image to the screen does not reliably improve learning (his "image principle"). The same research finds that an on-screen character who gestures and points at the content can help. A face that only talks mostly takes up room.
Screen space. A presenter takes a large share of a 16:9 frame. On a phone, the diagram left beside them is often too small to read.
Updates. When a policy changes, an avatar video needs a new script read to camera. An illustrated video needs the changed scenes redrawn, and the rest stays as it was.
Languages. A presenter speaking a translated script has to match lips to new audio. An illustrated scene only needs new narration and, if it has text, new labels.
Avatars still have their uses. For the bulk of a training library, they are the wrong default.
What replaces the face
Taking the presenter out only works if something better goes in. In an illustrated video, each scene shows what the narrator is describing at that moment: the step, the object, the before and after.
That is how Interlace works. You start from a document or a prompt, check the scene plan, and each scene is drawn in one visual style with narration in one of 28 languages. There are no avatars. The style does the job a presenter's tone would, so it matters which one you pick.
People still appear in illustrated video when the story needs them. A workplace scenario drawn in Ligne Claire or Graphic Novel shows a manager and an employee in a real setting, without an avatar reading lines at the camera. See all 18 in the video style gallery, or watch finished examples in the showcase.
Presenter or illustration: a decision table
Use this per video, not per library. Most teams end up with a mix.
| The video is... | Presenter (avatar) | Illustration |
|---|---|---|
| A CEO or leader announcement | Better | Works if there is no leader to cast |
| A personal welcome or sales message | Better | Weaker |
| A step-by-step procedure or SOP | Weaker: face competes with the steps | Better |
| A policy or compliance rule | Works for the intro | Better for the obligations and examples |
| A concept, system or data flow | Weaker | Better |
| Soft skills role-play | Better | Works with illustrated characters |
| Product or software training | Weaker | Better, alongside screen recordings |
| Needed in many languages | Needs lip-matched audio per language | New narration and labels per language |
| Updated often | New read per change | Redraw only the changed scenes |
A quick test: mute the video. If the learner can still follow the steps from the picture alone, the picture is doing the teaching. With a talking head, muting usually leaves nothing to follow.
Tools on each side, honestly
As of September 2026, two widely used avatar platforms describe themselves this way on their homepages:
Synthesia. Positions itself as an AI video platform for business, with more than 240 AI avatars and voiceovers in more than 160 languages. Choose Synthesia if a presenter on screen is central to your videos and you want a large avatar library. Our Interlace vs Synthesia comparison and the list of Synthesia alternatives go further.
HeyGen. Leads with lifelike avatars, including avatars of yourself, and video translation in more than 175 languages and dialects. Choose HeyGen if you want a digital version of a real person, or you need to translate videos you already filmed. See Interlace vs HeyGen.
Interlace. Choose Interlace if most of your videos teach a process, a policy or a concept, and you want to start from the document you already have. It has no avatars. You get illustrated scenes in one of 18 styles or a custom style from your brand, narration in 28 languages, captions, and MP4 or SCORM 1.2 export. It does not dub or lip-sync video you filmed elsewhere.
If you need both, that is a reasonable setup: an avatar tool for the handful of leader messages each quarter, and an illustrated tool for the procedure and policy library. Our AI training video software checklist helps you test either kind against your own documents.
Try it on a procedure
Take one SOP or policy your team trains on, and ask what the learner needs to see at each step. If the answer is never a face, you want an AI video generator without avatars. Join Interlace to make that video from your own document.