How It Works

Not a sentiment score. Not a clip list.

A structural framework for reading coverage at scale.

Traditional media tools count mentions and measure sentiment. Narrative Prism reads coverage the way a working analyst would: extracting the claims being made, the frames doing the work, and the voices doing the talking. Then tracking all three across the ecosystem in real time.

The Framework

Three components. Every story.

Narrative Prism decomposes every piece of coverage into its structural elements. The framework is consistent across topics, outlets, and time. That’s what makes coverage comparable in the first place.

01 · Claims

What’s being asserted.

The factual assertions a piece of coverage rests on. Extracted, attributed, and tracked across the ecosystem so contested claims can be seen for what they are: which sources are repeating them, which are pushing back, which are quietly dropping them.

02 · Frames

How it’s being told.

The interpretive structure beneath the surface of the words. Which conflict is foregrounded. Which actors are cast as protagonists. Which moral logic the story is asking you to accept. Frames are where narrative power actually lives.

03 · Voices

Who is speaking.

Full attribution by outlet, byline, and quoted source. Who’s amplifying whom. Which voices are gaining ground across the ecosystem. Which voices are being recycled across outlets that pretend to disagree.

Not a sentiment scoreNot a clip listNot a blended summaryNot a keyword alert
How It’s Different

Traditional monitoring counts coverage. Narrative Prism reads it.

Media monitoring was built to track mentions and measure sentiment. That was the right tool for a different era. The problem now is structural: knowing not just what’s being said, but who’s saying it, how it’s being framed, and which narratives are gaining ground.

Traditional MonitoringNarrative Prism
OutputArticle listsNarrative competition mapped across coverage
AnalysisSentiment scoresFrame analysis: how a story is being told
Attribution“Sources say”Full voice attribution by outlet, journalist, and quoted figure
AlertingKeyword alertsNarrative shift detection. Alerts fire when framing changes, not when words appear
SummaryBlended summariesCompeting narratives separated and tracked independently
A Working Example

What narrative competition looks like in practice.

Two competing narratives on a single topic, tracked across coverage. Source counts, key voices, frame divergence, all visible in one view. This is one of many topics actively tracked in the platform.

AI Regulation / Narrative Competition
Live
47
Articles
23
Voices
18
Outlets
+4 voices / 24h
Velocity
Narrative Tension78% · Escalating
Narrative A

“Regulation is necessary to prevent AI harms.”

26 Articles12 Voices9 Outlets
Key Claims
  • Existing harms (bias, surveillance, displacement) are documented and ongoing
  • Self-regulation has failed at every prior technological inflection point
  • Federal frameworks lag enforcement by years; harms accrue in the gap
Key Voices
Sen. MarkeyTristan HarrisTimnit GebruEU Commission
Carried By
NYTWaPoThe AtlanticGuardian
Narrative B

“Regulation will stifle innovation and cede ground to China.”

21 Articles11 Voices9 Outlets
Key Claims
  • U.S. AI leadership is a strategic asset that premature rules will erode
  • Compliance burden falls hardest on startups, entrenching incumbents
  • Adversaries face no equivalent constraint and will close the capability gap
Key Voices
Marc Andreessena16zSen. CruzPalmer Luckey
Carried By
WSJFox BusinessFree PressPirate Wires