RSV Global / Work / AI Economy Ledger
Built against one failure mode:
incomplete evidence looking complete.
A source-linked intelligence platform for the AI economy, built against a specific failure mode: presenting incomplete evidence as though it were a complete market model.
By the time a number is a headline, its context is gone.
Market analysis loses its source context, its coverage limits and its revision history somewhere between the research and the headline. By the time a number is quoted, the reader cannot tell what it excluded, when it changed, or whether the gap in the data was material.
Provenance
Every figure has to remain traceable to what produced it.
Coverage honesty
Partial coverage must stay visibly partial.
Rights
Not everything ingested can be republished, and that is a per-source question.
Reproducibility
A number cited last month has to be recoverable this month.
How the system is put together.
The choices that shaped it.
Each of these was a decision with an alternative, taken deliberately.
Missing evidence stays missing
No synthetic total is created because a complete-looking number is easier to present. This is the rule the whole product exists to enforce.
Four separate states, not one flag
Whether we hold the rights, whether it has been reviewed, whether it may be published and where it came from are four different questions. Collapsing them into verified loses the ones that matter.
Downloadable records
If a figure cannot be taken away and checked, it is a claim rather than evidence.
Edge deployment with scheduled monitors
Database, queue, ingestion and storage health monitors run on a schedule against production, because an ingestion failure that nobody notices becomes a silent coverage gap.
Where AI is used, and where it is not.
- Extraction and classification from source material.
- Drafting narrative around published figures.
- Producing a total.
- Filling a coverage gap.
- Deciding publication eligibility.
- Assigning rights or review state.
What it takes to keep it running.
Deployed to Cloudflare Workers with R2 object storage and OpenNext, split across public and private repositories, with separate operator configuration and scheduled production health monitoring across database, queue, ingestion and storage.
What it changed.
Measured outcomes for this system are being compiled from our own operating records. We publish figures only once we can show how they were measured, so this section is deliberately empty until then.
More systems we built and run.
Show us the workflow everyone has learned to tolerate.
We will help determine whether it needs a focused product, an autonomous operating system, a stronger data layer, a review of what already exists, or a simpler fix that involves no AI at all.