MOTHERSHIP

[work] ai newsroom

Public TV.

An autonomous newsroom.

Public TV's pipeline was not broken. It was built, professionally, for the 8pm bulletin, while the audience moved to feeds that never sleep. Their question: what can AI actually do here? Our answer: an autonomous newsroom, from live feeds to published clips, with one operator supervising.

editorial agent crewself-hosted talking headsUE5 MetaHumansoperator cockpit

THE FLOOD

feeds · articles · social · wire · clips. Stories arrive faster than humans can package them

THE NEWSROOM · AGENTS

director · curates beats from live feeds
researchers ×3 · parallel deep-dive
writer + stage director · script, shots, pacing
editor + sponsor · review, ad placement
one operator supervising every agent from Mission Control

THE STORY PACKAGE · MADE ONCE

script · voice · lipsynced presenter · branded composite

ON AIR, EVERYWHERE

shorts · bulletins · long-form · sponsor slots, published where the audience actually lives

[01] the bottleneck

The feed never stops. Production does.

Five to eight people in sequence. 45 to 120 minutes per clip. Most of it assembly, not journalism.

A media company with a decades-deep broadcast operation came to us with a sharp brief: how do we switch up the social media game, and what can AI actually do here? Nothing about their pipeline was broken; it was built, professionally, for a different world. The 8pm bulletin, the studio floor, the broadcast tower.

We started where we always start: how the work actually happens. A short-form clip touched editor, researcher, writer, producer, voice, video editor, motion designer, and reviewer in sequence. The stories aged faster than the pipeline could package them. That is not a talent problem. It is a shape problem.

[02] editorial agents

Editorial agents that find the beats.

Ingest -> Enrich -> Curate -> Script -> Voice -> Presenter -> Composite -> Review -> Package.

The Hot Take loop: a Director agent discovers and curates stories near-instantly from live feeds. Parallel Researcher agents deep-dive the strongest candidates. A Writer turns research into script, a Stage Director adds shots and pacing, an Editor reviews, and a Sponsor agent reads the finished piece and places contextual ad segments at natural breaks. Mission Control shows an operator what every agent is doing.

09:04 AM · A STORY BREAKS
09:04

Feed item lands

“Metro Line 3 opens six months early after final safety clearance...”
live feedunverified
09:05

Director agent curates it

Scores the beat against editorial memory: covered twice before, this is the payoff development.
beat: metro line 3format: Reactionpriority: high
09:07

Researchers deep-dive, in parallel

Article extraction, prior coverage, social reaction, official statements. Context assembled while the story is still news.
14 sources3 agents
09:12

Written, voiced, performed

“Six months early. Three stations. And the first trains roll tonight...”
Anchor script → designed voice → ultrarealistic lipsynced presenter, rendered on self-hosted GPU infrastructure.
09:17

Composited to channel grammar

Lower thirds, ticker, channel bug, vertical format. It looks like the channel, because it is the channel.
09:22

The human moment

The operator checks sourcing, accuracy, and brand. Judgment, not assembly.
PublishWatchBundle
09:24

On air, where the audience lives

YouTubeInstagram+ story package saved for the bulletin, the deep dive, and the sponsor slot
THE OUTPUT
PTV
METRO LINE 3 OPENS SIX MONTHS EARLY
First trains tonight · three new stations
BREAKING · METRO LINE 3 CLEARED FOR SERVICE · FARES UNCHANGED THROUGH MARCH · NEXT: WEATHER
Ultrarealistic lipsynced anchor · branded compositing · vertical + broadcast formats from one package

MISSION CONTROL

Director · curating live feedsdone
Researchers ×3 · parallel deep-divedone
Writer · anchor scriptdone
Stage Director · shots, pacing, overlaysrunning
Editor · review passqueued
Sponsor · reads final, places ad breakqueued

[03] presenter layer

Anchor scripts become anchor video.

Designed voices. Lipsynced presenters. Open-source models, self-hosted.

Ultrarealistic talking heads at channel-grade quality, running on open-source models we self-host on RunPod servers spun up for production. That is what makes the economics work at daily volume, where hosted alternatives price a single 30-second clip in the multi-dollar range.

When the format demands cinema, the same editorial brain drives an Unreal Engine pipeline: MetaHuman performance, automated sequencing, film-grade rendered output.

[04] the cockpit

Taste, quality control, human in the loop.

Every stage observable, editable, and re-renderable.

An operator cockpit where every stage can be inspected and re-rendered without restarting the job. Branded compositing so the output looks like it belongs to the channel, not to a template. Editorial memory underneath: beats tracked across days, story packages reused across shorts, bulletins, deep dives, and sponsor inventory.

[05] the outcome

The human moves up the stack.

One operator supervising instead of five to eight in sequence. Minutes instead of hours.

1operator supervising, not 5 to 8 in sequence
~20 minfrom break to broadcast
$1-2per finished clip, self-hosted
10+shorts a day from the same packages
daysof beat memory behind every story
every stageinspectable and re-renderable

An autonomous newsroom: catching beats as they break and publishing to Instagram, YouTube, and wherever the audience actually lives, with a minimal human in the loop. The human does not disappear; the human moves from assembly to judgment, sourcing, taste, and the publish decision. The economics are stated as the operating model the system is built to, not audited telemetry.

The machine generalizes. Sports highlights, market commentary, regional-language news, branded explainers: any operation where stories arrive faster than humans can package them, and the package has to look like it belongs on a specific channel.

[next] explore

Adjacent capabilities.

Powered by Broadcast and Workforce.