What is AI video, and where does it help a small business?
Written lesson · about 2 minutes to read
The short answer
AI video comes in three kinds: editing help on your own footage, video generated from a description, and talking avatars. For a small business the first is the most useful today, because it adds captions, cuts and clean sound to real footage you already have.
Three things to remember
- 01The useful kind today is editing help on your own footage: captions, cuts, cleaner sound and short clips.
- 02Never put words in a real person's mouth, and get consent before using anyone's likeness or voice.
- 03Tell viewers when footage is generated rather than filmed.
Three kinds of AI video
The first kind is editing help on footage you filmed: automatic captions, cutting out pauses and mistakes, cleaning up background noise, and pulling short clips from a long video. The second is generated video, where you describe a scene and the tool makes footage that was never filmed.
The third is the talking avatar, a digital presenter who reads a script on screen, sometimes in a cloned voice. These are three different tools with three different risks, and only the first works on something you actually filmed.
Why editing help is the useful one
It works on real footage: a phone clip of a finished job, your senior tech explaining a fix, the owner answering the question customers ask most. Nothing false is created. It is still your work and your voice, only tidier and shorter.
Captions matter because many viewers watch without sound. A long recorded talk can become several short clips for social posts. Read the captions before you publish, because names, part numbers and trade terms are where they go wrong.
Where generated video and avatars need care
The rule from images applies here too. Do not use generated video to show your work, your premises or real people. Never make video of a real person saying something they did not say, and get their consent before using anyone's face or voice, whether that person is a member of staff, a customer or the owner.
Tell viewers when footage is generated, in a plain line on the video. Trust is the reason. Rules on using a person's likeness and voice differ by state and country, so check with a lawyer for your own case.
A talking avatar may suit an internal training video, but customers often find a digital presenter cold. If you use one, say so, and never build one from a real person's face or voice without their consent.
Where to start this week
Film one short video on your phone: a finished job, with the customer's permission, or a tech answering the question customers ask most. Ask the person on camera whether they are happy for you to post it.
Then use an editing tool with automatic captions, cut the pauses, check every caption and post it. Next time, ask the tool for two or three short clips from a longer video. Leave generated video and avatars until the editing habit is working.
More in part 7
- 1Can AI answer your phone?
- 2What is speech-to-text?
- 3What is text-to-speech?
- 4Can AI transcribe and summarise your calls?Written
- 5Can AI translate for your customers?Written
- 6What is document parsing (OCR)?
- 7How does AI read invoices, purchase orders and forms?Written
- 8Machine vision vs. AI vision
- 9What is AI image generation, and can you use it for marketing?Written
- 10What is AI video, and where does it help a small business?WrittenYou are here
Next step
Where does this fit your business?
Six quick questions about your business, then your AI plan: what to start with, what a person still checks, and the lessons that go with it. Or bring it straight to a free thirty-minute call.