LinkedIn Will Now Cut Your Event Video For You.

LinkedIn has added AI recommended clips and chapters to live event video recordings.

The composer reads the recording, flags the moments it identifies as major announcements, key insights or memorable quotes, and hands them back as suggested segments the host can shorten or extend. Chapters are generated inside longer recordings so a viewer can skim a replay rather than scrub through it.

LinkedIn says recommended clips and chapters are available to all users globally.

So the editing cost of turning an event into short video has largely gone. Editing was never the thing stopping brands doing it.

What LinkedIn actually shipped

The recommended clips composer sits inside the event management tools. Instead of an empty timeline, the host gets a panel of suggested moments and picks from it, adjusting the in and out points of each one.

Chapters are the second half. On a longer recording, the segments are labelled so someone landing on the replay a fortnight later can find the part they came for.

Worth being precise about what is new. LinkedIn already listed a clipping tool on its marketing product updates page in March 2026, described as turning an event recording into LinkedIn ready video clips by selecting key moments. The selecting was yours. Now the first pass at selecting is the model’s.

Editing was never the bottleneck

Ask a marketing lead why last quarter’s webinar produced no clips and you will not hear “we could not cut it”. Editing an event down is a solved problem, and has been for years.

What you hear instead is that nobody watched the recording back. Or that someone did, sent four timestamps to an editor, and the file sat in review until the topic went cold.

The cost was never the cut. It was attention, ownership and a decision. LinkedIn has automated the cheapest step in the chain, which is genuinely useful and does not touch the expensive one.

Why a model picks the moment that sounds best

A recommendation engine looking for memorable quotes optimises for quotability. That is a real signal and it is not the same signal as buying intent.

The most quotable forty seconds of a webinar is usually the confident generalisation near the top. The line that actually moves a buyer tends to be the awkward middle, where the speaker answers the objection everyone in the room was carrying.

That moment rarely sounds polished. It has a hesitation in it, a caveat, a number the speaker had to look up. Those are the properties that make it credible, and they are exactly the properties an engine trained on punchiness will score down.

This is the same reason clipping is not the same as repurposing. A clip is a slice of a recording. A post is an argument that happens to be delivered on video.

Selection is the job now

If the model does the cutting, the remaining human work is deciding which suggestions are worth publishing, in what order, and what each one is for.

That work needs an owner. Not a committee, not a tool, one named person who watches the panel of suggestions and says yes to four of them.

In most teams that person does not exist, because the role was never needed while editing was the visible cost. The visible cost has moved, and the org chart has not moved with it.

Give it to whoever knows the sales conversation best. Someone who has heard the objections is a better selector than someone who has only heard the recording.

How to choose which forty seconds go out

Run every suggested clip past the same five questions before it becomes a post.

1. Does it answer a question a buyer actually asks, in words they would use?
2. Does it stand alone, or does it only make sense to someone who watched the hour?
3. Is there a specific in it, a number, a name, a constraint, rather than a general claim?
4. Would the speaker be comfortable defending this thirty seconds out of context, because that is how it will be seen?
5. Does it earn its place next to the other clips from the same event, or repeat one of them?

Anything that fails question two is not rejected, it is rebriefed. Most near-misses become publishable with a rewritten opening line and eight seconds trimmed off the front.

Build a ten minute review step

Selection needs a slot in the calendar or it will not happen. Book it while the event is still in the diary, not after.

A workable shape is one pass within forty eight hours of the event: the owner opens the recommended clips, marks the ones that clear the five questions, writes a one line note on each saying what it is for, and sends the set on.

That is a review step, not an approval queue, and the difference matters. Approval asks whether something is allowed. Review asks whether it is worth the slot. If your process only has the first, the approval process template is a faster starting point than inventing a lane from scratch.

Keep the loop tight. One reviewer, one round, a fixed turnaround.

Chapters are the quieter half

Clips get the attention, chapters do the durable work.

A replay with labelled sections is a page someone can be sent to. A sales conversation can point at the chapter that answers the pricing question rather than asking a prospect to watch fifty minutes on faith.

Check the generated labels, because they are only as good as what the model heard. A chapter called “Q and A” tells nobody anything. Rename it after the question it answers.

What good looks like

Suppose a B2B software brand runs a quarterly webinar. Good is not eight clips landing on the Monday.

Good looks like this:

• One owner who watches the suggested clips and picks, with the authority to say no to six of eight.
• Four to six published clips per event, each one attached to a question the sales team hears.
• A publishing rhythm that spreads them over the fortnight after the event rather than dumping them in a day.
• Chapters renamed by hand so the replay is navigable by a stranger.
• A short note after four weeks on which clip earned the most qualified conversations, feeding the next event’s run of clips.

Nothing there is about editing speed, and all of it is organic distribution of something you already made.

How NBK thinks about event repurposing

Every time a platform removes a production cost, the constraint moves somewhere less visible, and usually into the process. The teams that get value from this will be the ones who already know who decides what ships.

That is the whole question of who owns the creative brief turning up again in a new shape. A suggestion engine is a first draft with no point of view. Somebody still has to supply the point of view.

NBK’s position on tools like this is consistent. Let the platform take the mechanical step. Keep the judgement, the selection criteria and the review in your own system, because that is the part competitors cannot copy.

Next step

If your events keep producing recordings that nobody turns into anything, the fix is not a better editor. It is naming who selects, what they select against, and when.

If your social process is slowing down good ideas, NBK can help rebuild the workflow behind the content.