You already wrote the post. Now say it out loud.

The phrase has been worn out by people who post quotes over stock sunsets, but the thing underneath it is simple: a specific person saying a specific thing is what stays with people. The reason so much of it reads like nothing is that it was written to be inoffensive.

Three rules that hold up.

• One idea per post. If the second paragraph starts a new argument, that is next week's video. Two arguments are two posts, and both land better split.
• The first line does the work. Someone scrolling decides in a line or two, long before they reach your point. Start on the claim, not on the setup. "Most founders hire a salesperson a year too early" beats "I have been thinking a lot about hiring lately".
• Say the inconvenient one. Say the true thing, the one you would say off the record. You almost certainly have a view on your own industry that you have never posted because it costs you something to say it. That is the one.

If you want help getting from a rough thought to a run of posts, Kae plans it with you before anything renders and shows you the cost first. She is our creative director, and she is on every plan, with 5 Kae planning sessions a month on Free.

Plenty of people scroll with the sound down, in an open-plan office or on a train, and they will read your video before they ever hear it. Two things follow from that.

Front-load the claim. Whatever you want understood goes in the opening line of the post text and in the first sentence you say, before any build-up.

Burn the captions in. Here is the exact route, because it is not one button. Download your render, open the Editor, start a new edit, and drop the file in. Add the captions there. They are drawn into the picture rather than shipped as a separate subtitle file, so they cannot be switched off, lost on re-upload or stripped when someone downloads it. The export bills 20 credits a second, with a floor of 60 credits, so a short cut is a small, predictable cost on top of the render.

On LinkedIn these are usually two different people. The partner has the profile. The chief of staff has the calendar, the drafts and the posting app. That split is fine, and it is worth being precise about how it works here.

A Digital Twin can only exist because the person it is built from recorded their own consent clip. Nobody can do that step for somebody else, and there is an identity check on the recording before the twin can render anything. Once that is done, the twin lives in your Workspace and whoever you invite can write and render with it. The person whose face it is can withdraw that consent later, and then it stops.

So the honest description of the working arrangement is this. The principal gives about two minutes, once. Everything after that belongs to whoever writes for them. If your firm has several partners who post, each one records their own clip and gets their own twin. Plus includes 3 character slots, Pro 10 and Max 25, and a Digital Twin and an Avatar draw from the same pool, so a made-up presenter you build for something else uses one of the same slots.

You are publishing a video of your own face, under your own name, to people who know you. The way to keep that trust is to be boring and direct about how it was made. One line at the end of the post does it.

This video was made with an AI version of me, using my voice, from a script I wrote.

That is the whole disclosure. It does not need a paragraph of reassurance, and it reads better than being asked about it in the comments later. Built-in AI labelling is still rolling out across output types here, so write the line yourself rather than assuming the file carries it.

Two things not to do. Do not build a twin of someone who has not recorded their own consent, including a colleague who said yes in a meeting. And do not put words in your own twin's mouth that you would not sign your name to in text, because on LinkedIn your name is already on it.