How Fanatics Is Building AI Into the Way It Operates

Fanatics runs 100+ livestreams at once. Games, drops, and shop-alongs move in real time, and the moments that sell a card last seconds. The work is not watching every hour. It is catching the play, the pull, the reaction, then having a clip ready to post.

That is what Auto Clipper does on fanatics.clippable.io. It sits on those ongoing livestreams and pulls the card moments from each one, so operators review clips instead of sitting through the stream.

From watching streams to a clipping desk

A single good clip is not the hard problem. The hard problem is running the same job across 100+ livestreams without a new brief every time. Auto Clipper is that desk: one workspace, every live feed, card moments in posting order.

Each stream is a long video that never really ends. Auto Clipper ranks the important windows: the card hit, the crowd, the reveal, and cuts shorts with captions. People spend time on which clip ships, not on scrubbing.

Automating recurring work

Card moments repeat: a hit, a pull, a crowd. Auto Clipper treats that as a loop, not a one-off edit from a blank timeline.

The same desk watches every live feed. The clips come back already cut:

Auto Clipper watches the stream. It pulls the card moments. People still pick what ships.

Operators are not collecting source hours. They are choosing among moments Auto Clipper already pulled from the livestream.

Bringing more context to live decisions

When a card moment happens, the window to post is short. Auto Clipper brings the cut to the operator while the stream is still live.

It does not decide the merch story. It gives them the moment before the window closes: which hit, which angle, which caption still belongs on the card.

The bottleneck is not whether the clip exists. It is whether the card moment reaches the operator before the stream moves on.

The difference from a faster editor is simple: the output is no longer a file from last night’s VOD. It is a clip from a livestream that is still on.

Building around the workflow itself

The workspace is fanatics.clippable.io, not a new tab for every stream. Clips land in the Auto Clipper desk. Review, recut, or queue.

Nothing posts until someone says so.

AI prepares the cut. People approve. Then it can post.

Related streams sit in the same desk instead of a pile of separate recorders.

The live feeds, the card moments, and human approvals wrap around a recurring post.

A loop centered around human judgment

Auto Clipper is not replacing the people who know the card. It moves them to the moments that matter: which hit is the post, which caption is honest, when to ship.

As streams keep running, prior approvals can inform the next cut. That is the operating loop: Auto Clipper watches the livestream, people make the call, and each card moment can inform the next one.

For a business that moves with the game, the advantage is not watching more stream. It is posting the card moment while it still matters.