Try every tool free for 14 days with 250 credits. No card needed. Start free
Product
AI video generatorText or an image to video with Veo, Kling, Seedance and more AI avatars & lip syncTalking presenters and lip-synced video from a script or audio UGC video adsAn AI creator presents your product on camera, in 9:16 Subtitles & translationTranscribe, edit, translate and download SRT or VTT Podcast to clipsTurn a long video into captioned vertical clips AI dubbingDub a video into 32 languages, in the speaker's own voice AI voice & cloningNatural voiceovers and your own cloned voice
Company
Free tools Pricing Blog Contact

How to Make a Training Video (Step by Step)

A practical process for employee, customer and onboarding training videos: planning, scripting, recording the screen or using an AI presenter, captions and translation, and updating it later.

How to Make a Training Video (Step by Step)

To make a training video, choose one task the viewer should be able to do afterwards, write a short script at a slow pace (120-140 words per minute), record the screen or a presenter, add captions, and test it on someone who has never done the task. Keep each video to one task and a few minutes; a course is a series of short videos, not one long one.

More on this topic: Video marketing, every guide in one place.

This guide walks through the seven steps for software, process and customer training, with the free and paid tools that fit each, and how to keep videos up to date when the product changes.

The short answer

1) Pick one task and the result. 2) Write a script of 360-720 words for a 3-6 minute video. 3) Choose the format: screen recording, talking presenter, AI avatar or slides with narration. 4) Record with a quiet mic, or generate the voice. 5) Edit out dead air and add on-screen labels. 6) Add captions, and translations if needed. 7) Test it on a new user, then publish where learners already are.

Key takeaways

  • One video, one task. Title it with the task (“How to approve an expense”).
  • Script at 120-140 wpm: 3 minutes is about 360-420 words.
  • Screen recordings are the fastest format for software; presenters help for policy and culture topics.
  • Captions are essential: many people watch training on mute, and some need them.
  • Plan for updates: short videos and scripted voiceovers are cheaper to redo.

Disclosure: we make QuickVeed and show how to do this in it, along with other ways. Details about other products come from their public pages, checked in October 2026 and linked at the end.

Step 1: Decide what the viewer must be able to do

Write one sentence: “After this video, a new team member can …”. If you need two sentences, you need two videos. Then list the steps exactly as someone would do them, and note the one or two mistakes people usually make; those deserve a callout.

Step 2: Write a short script

Even screen recordings go better with a script. Aim for 3 to 6 minutes per video:

LengthWords at 120 wpmWords at 140 wpm
2 minutes240280
3 minutes360420
5 minutes600700
6 minutes720840

Training is slower than marketing, because viewers have to follow along. Paste the draft into the free words to minutes calculator to check its length, and see average speaking rates and words per minute of speech for pace. Structure: what we'll do (one line), the steps, the common mistake, and where to get help.

Step 3: Choose the format

FormatBest forTools
Screen recording with voiceSoftware, internal tools, customer onboardingFree screen recorder; built-in recorders
Talking presenter (real person)Policies, culture, leadership messagesFree webcam recorder and teleprompter
AI avatar presenterShort intros and recaps you'll update or translate oftenQuickVeed AI avatars, up to 30 s per video
Slides with narrationConcepts, compliance, numbersPresentation app + voiceover
MixedMost coursesA presenter intro, then the screen

For software, start with a screen recording: how to screen record on any device covers Windows, Mac and Chromebook. For presenter-led training at scale, AI avatar training videos explains where avatars work and where a real person is better, and HeyGen alternatives and Synthesia alternatives compare dedicated training tools (some export SCORM for an LMS, which QuickVeed doesn't).

Step 4: Record, or generate the voice

  • Sound first. A headset or USB mic in a quiet room beats a better camera. Viewers forgive a plain picture, not bad audio.
  • Screen: record one window at 125-150% zoom, hide notifications, and use test data, never real customer records.
  • Presenter: read from a teleprompter at eye level; QuickVeed's free teleprompter scrolls your script in the browser.
  • Voice without recording: QuickVeed's AI voice generator reads up to 10,000 characters a job with 9 models, from 7 credits per 1,000 characters. Re-voicing an updated step is a text edit, not a re-record.

Step 5: Edit for clarity

  • Cut the dead air at the start and end with the free video trimmer.
  • Add a short label for each step (“Step 2: Approve”) with the free Add Text to Video tool.
  • Zoom or crop to the part of the screen that matters; the free video cropper helps.
  • Speed up waiting (a page loading) rather than leaving it in.

Step 6: Add captions and translations

Captions help people watching on mute, in a noisy place or in a second language, and many organisations require them for accessibility. QuickVeed's auto subtitles transcribe the video, let you correct every line (product names especially), export SRT or VTT for your LMS or video host, and burn captions in one of 11 styles. A transcript costs 1.5 credits a minute, rounded up, and a burned-in copy 1 credit a minute with a 3-credit minimum, so captions on a 1-minute clip cost 5 credits. Translations into 32 languages keep each line's timing.

For a voice in another language, QuickVeed's AI dubbing re-voices a video with stock voices; AI dubbing vs subtitles helps you choose, and how to lip sync a video covers matching a presenter's mouth to the new audio.

Step 7: Test, publish and keep it current

  1. Watch a new colleague follow the video without help. Where they pause or ask, fix the video.
  2. Publish where people already learn: your LMS, help centre, intranet or an unlisted YouTube playlist.
  3. Put the date and product version in the description, and review the set every quarter or after each release.

Name files and titles by task so people can search them, and link related videos at the end of each one.

Frequently asked questions

How long should a training video be?

Usually 2 to 6 minutes, one task per video. Split longer topics into a series.

What is the best format for software training videos?

A screen recording with a clear voiceover and step labels, plus captions.

Can I make training videos with AI?

Yes: AI voices for narration, AI avatars for short presenter segments, and auto-captions and translation. A human should still check every step and term.

Do training videos need captions?

They should have them. Many learners watch on mute, and captions support accessibility and non-native speakers.

How many words is a 5-minute training video?

About 600 to 700 words at a training pace of 120-140 words per minute.

How do I keep training videos up to date?

Keep them short and scripted, so you can re-record or re-voice one step when the product changes.

Try it with your own video

Record the steps for free, then add a narrated intro, captions and translations in QuickVeed. Upload your screen recording, check the captions line by line, and judge them for yourself. QuickVeed plans start at $18/month, and new accounts get a 14-day trial with 250 credits, no card needed. Sign up or see pricing.

Share this article
Ijharul Islam
Written by

Founder, QuickVeed

Ijharul Islam is the founder of QuickVeed, the AI video studio for business. He writes hands-on comparisons and guides on AI video generation, presenters, UGC-style ads and subtitles, so marketing and sales teams can publish more video without a production crew.

Chat with the founder