Employee onboarding video: a new-hire series made with AI

An employee onboarding video is a short video new hires watch in their first days to learn who the company is, how work gets done, which tools to use, and who to ask for help. It works best as a series of short clips, one topic each, so you can replace a single clip when a policy or tool changes instead of reshooting everything.

With an AI avatar, each clip is rendered from a script, so an update is a new render rather than a new shoot. On Sume, each clip is one talking-video job that reuses the same avatar handle. Facts come from Generate avatar video and the Models overview, read on 2026-09-28, plus Sume's current code where noted.

Which videos should an onboarding series include?

Plan the series as a list of topics, not one long film. A common set looks like this:

• Welcome: who the company is and what the first week looks like.
• How we work: meetings, working hours, and how decisions get made.
• Tools: which app is used for what, and where to request access.
• Who to ask: the people or channels for IT, HR, and payroll questions.
• Team clips: one per department, owned by that team's lead.

How do I make onboarding videos with an AI avatar?

• Create one avatar and keep its handle. The docs recommend a stable handle so your app reuses the same presenter; creating a reusable AI avatar shows the call.
• Write each topic to fit one job. A talking video is accepted when Sume estimates it at 4-60 seconds; split a longer topic into two jobs.
• Pick the shape. aspect_ratio supports 1:1, 3:4, 9:16, 4:3 and 16:9, and the default is 9:16, so ask for 16:9 if new hires watch on a laptop.
• Set the room with scene: { "type": "prompt", "prompt": "..." }, for example an office.
• Send an Idempotency-Key header so a retry after a timeout doesn't start a second paid job.

curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: onboarding-tools-v1" \
  -d '{
    "avatar_handle": "onboarding_host",
    "script": "Welcome to your first week. Here is where to find the tools you need.",
    "aspect_ratio": "16:9",
    "scene": { "type": "prompt", "prompt": "Bright open-plan office" }
  }'

How do I update one video when a policy changes?

Edit that clip's script and submit a new job with the same avatar_handle, aspect_ratio and scene. Use a new idempotency key: the docs say to reuse a key only for the same operation and payload. The other clips stay as they are.

Completed results can include public media.sume.com video files, and the docs tell integrations to store the Sume URL. Replace the old clip's URL wherever it is linked. To check the framing before paying for a full render, an avatar video preview returns first-frame stills first.

Can new hires watch it as one long video?

Yes, if you join the clips yourself, because one job stops at about 60 seconds. Timeline 1.0 takes one audio spine plus ordered video slots and returns one MP4 of up to 1,800 seconds. In current code a Timeline render's sound comes only from the spine and an optional soundtrack, so each clip's speech has to be carried on the spine. Making an avatar video longer than 60 seconds walks through the join. Separate clips are easier to replace.

What are the limits, and what does it cost?

• English only: in current code the avatar route speaks English. What languages can an AI avatar speak covers the route for other languages.
• One presenter per clip: current execution supports one resolved avatar per final video.
• 720p: resolution is currently 720p.
• No screen recordings: there is no public upload route for local files, so a software walkthrough needs a screen-capture tool.
• Hosting is yours: Sume returns MP4 files. Converting a video to SCORM covers LMS packaging.