storyobjectmodel.comdemos

Two breaking news scenarios. One story bus.

SOM does not tell you what to build or how you and your technology partners implement AI. It enables; it does not define. So the consortium built two realistic newsroom scenarios with different stories and different vendors, all reading and writing the same story on one bus. Old tools and new. AI and not. Nothing in either scenario sends a command to anything.

scenario 01

The hurricane

A storm becomes a hurricane, with one report nobody can confirm. A real wire corpus and a weather data feed on the bus, transcription and media intelligence reading the story. Two clips arrive: one carries its credentials, one does not survive the gate, and the record is kept. The classification changes at source, a journalist confirms it, and everything that depended on the old version finds out in the same second. A single-source claim stays off the page while the page publishes: a system understood something well enough to do nothing. Then the briefing lands, the transcript confirms it, and the clearance becomes a record on the story that every tool can see.

AP · The Weather Company · Trint · Moments Lab · Cuez · Vizrt · HyperContent AI · Shure · Amazon Web Services
APThe Weather CompanyTrintMoments LabCuezVizrtHyperContent AIShureAmazon Web Services
scenario 02

The Prime Minister announcement

A statement outside Number Ten, and nobody knows what he will say until he says it. The newsroom writes the story to the bus with the premise it expects and two sets of scripts, neither preferred. Tools reading the story draft and hold social posts, suggest a live unit, transcribe the feed and propose the claim the moment the words are spoken. Nothing changes until a journalist approves it. One approval settles three things, and then four things move: the clip is cut, the right scripts commit to air, the wrong ones are suppressed with the reason recorded, and the held posts release. Nobody was told. They were all reading the same story.

Reuters · Octopus · LiveU · Fonn Group (Mimir) · Trint · HyperContent AI · Amazon Web Services
ReutersOctopusLiveUFonn GroupTrintHyperContent AIAmazon Web Services

What both prove

Five things, proved twice, with different stories and different vendors.

ProvenanceThe evidence travels with the story. Where a clip came from, who vouched for it and what was done to it arrive with it, rather than in somebody's memory of the afternoon.
AccuracyThe premise changes once, and every dependent is flagged. One correction at source, and nothing downstream is left quietly running on the old fact.
ReachClear once, use everywhere. A decision made for one output is available to all of them, and a rewrite reopens the gate rather than inheriting the approval.
ComplianceA held figure reaches no output until it is cleared. The record says who cleared it, when, and against which rule.
SynchronisationNobody synchronised the tools. No orchestrator, no integrations between them. They were all reading the same story.
read the specification