CONTACT / Sales + partnerships

Want to solve a financial problem
with Treasurer AI?

ESTIMATE / Project estimate

Request a project estimate

Story/Media
STORY · MEDIA

Faster than breaking news:context, already organized

Reporting, research, story production — scattered filings, news and data gathered into one layer, with the tools your journalists use built on top.

01
AGENTS
Agents built
02
MODULES
Modules composed
03
PACKAGE
Package complete
AGENT · A1
News Collector
AGENT · A2
Filing Analyst
AGENT · A3
Company Research
AGENT · A4
Article Writer
AGENT · A5
Report Builder
MODULE · M1
News Intelligence
Company news collection
Industry trend monitoring
1 AGENTS
MODULE · M2
Company Research
Filing analysis
Company analysis
2 AGENTS
MODULE · M3
Report Generation
Article drafting
Report generation
2 AGENTS
PACKAGE
READY
Market Intelligence
News Intelligence
1 agents
Company Research
2 agents
Report Generation
2 agents
3 MODULES5 AGENTS
Groups into
Assembles
EXAMPLE

Reporting, production, publishing — every screen
a newsroom uses, on one data layer.

Trading Desk

System integration, live monitoring, signal detection — a trading desk’s day on a single screen.

  1. 01System integration
  2. 02Live portfolio monitoring
  3. 03Trading signal detection
System integration3 / 3 connected

PMS, OMS and trade data connected into one working environment.

HOW WE BUILD IT

We build the data first, not the tool.

Most solutions put a screen on top of someone else's data and models. We own the data layer, and build your newsroom's tools on top of it.

0.1 · DATA LAYER · OWNED OUTRIGHT

We do not borrow data — we own it

Filings, news, prices and archives collected and refined from the source into a single layer we own. Every reporting tool runs on our data, not licensed data.

0.2 · ONTOLOGY · STRUCTURE

Unstructured data into a digital twin of the newsroom.

People, companies, events and statements tied together in a news ontology — an archive asset that deepens as it accumulates.

PR · #284ready
▸apps/
▸pms/
·rebalance.pyM
·policy_gate.pyM
▸tests/
·test_gate.pyM
▸lineage/
·writer.py
+146 / −38 · 4 files
harden policy gate · raise cosign pathsha 6c91…ab
12@@ apps/pms/policy_gate.py @@
13 def evaluate(plan, ctx):
14- if plan.notional > 5_000_000:
15- raise PolicyBlock()
16+ if plan.notional > ctx.limits.notional:
17+ return Gate.require_cosign(plan)
18 return Gate.ok()
ci · passcoverage · 94%risk · low
Reviewers · your side first
Y
yujin · quant
owner
✓
M
minho · pm
code
✓
T
treasurer / ax
pair
✓
Squash & merge · your hand on the trigger
0.3 · REUSE · 10× FASTER

Once the first AX lands, everything after is 10× faster

We do not rebuild reporting, desking and publishing workflows from scratch each time. Modules and ontology built once get reused, so new tools ship 10× faster than conventional development.

0.4 · SECURITY · COMPLIANCE

We meet your security and compliance standards as written

Private VPC, isolated deployment and network separation put it safely inside the newsroom network. Software standards through to data-privacy requirements — every internal compliance and security bar met.

A DAY AT THE DESK

A day in the newsroom,
screen by screen

KST · publishing day
07:00
Morning items
AI summarizes overnight filings and foreign news
09:30
Intraday detection
Live alerts on sharp moves and unusual filings
11:00
Reporting research
Cross-checking archive against data
14:00
Story drafting
Review the AI draft · verify sources
16:30
Desking
Fact-check · automated copy-editing pass
18:00
Publish & report
Published automatically · performance tallied
INSIDE ONE FIRM

Not eight tools.
Eight modules inside one newsroom.

A single financial media outlet runs every module below — reporting, production and publishing do not drift apart. Same data, same sources, same verification record.

01
One data layer

Every module runs on filing, price and news data we collect and own. No dropped external feeds, no licensing gaps.

5 countries · owned outright
02
One lineage

Every number in a story traces back to its origin — filing, price, press material. You can reopen the same source a year later.

Event → action · hash
03
One signed ledger

Desking, edits and corrections all land in a single ledger. Reproduce a story exactly as it stood at any moment, and export the evidence bundle in one go.

Sign · replay · export
8 modules · 1 desk
01
Live item detection
Sharp moves · unusual filing alerts→
02
Filing source reading
DART · key sentence extraction→
03
Story archive search
Past coverage · context links→
04
Fact-check assist
Figures · quote source matching→
05
Story draft generation
Source citation · lede suggestions→
06
Desking & copy-editing
Style · automated sentence review→
07
Publishing & distribution
Per-channel publishing · status tracking→
08
Performance reporting
Views · referral traffic, tallied automatically→

Design the package that fits your firm,
together with Treasurer.

Through a free PoC or a project quote, we propose the AX direction that suits your organization.