Provenance-native intelligence for capital markets

Decide on data you can defend.

An intelligence layer that enriches any entity — a deal, an asset, a sponsor, a borrower, a counterparty — where every datum arrives carrying its lineage, confidence, and information-barrier status. By construction, not by clean-up.

Beachhead: real-asset & SMSF diligence — where the data moat, the licences, and the Cotality / RP Data backbone already live.
Friction here is a feature: access is granted, not self-served. We meet the bar institutions are required to hold.

Now forming Design-partner cohort — real-asset originators & SMSF diligence desks. [[ N seats ]] Apply to the cohort →
The reframe

Built for coverage.
Repurposed for proof.

Sales tools optimise for reach and speed — provenance is exhaust, because no one asks a rep to defend an email's chain of custody. Capital markets inverts every term. The plumbing of lead-gen is the product here.

The growth tool does…
…the capital-markets layer does
Coverage & speed"in minutes," "every account"
Provenance & defensibilitya citable trail per data point
Lineage is exhaustdiscarded after enrichment
Lineage is the assetsource · confidence · freshness · cost · lane
A clever firewallan engineering convenience
A regulated information barrierthe construct finance must maintain by law
Calibrated guessesresearch nudges for a rep
Citable diligencean analyst that refuses to assert the unsourced
Enrich a person to sella row in a list
Enrich an entity to decidea deal, asset, sponsor, counterparty
Getting caught = account banrecoverable
Getting caught = regulatory eventtrust you can't re-earn — so we never risk it

Same primitives. Opposite gravity. In capital markets, rigour is paid for and provenance is mandatory.

What "by construction" means

Proof is a property of every cell.

Not a report you assemble at the end — a guarantee the data structure carries from the first field.

1 : 1
Every datum bound to its source. No unattributed values reach the clean lane.
5-axis
Each cell tags value · source · confidence · freshness · cost — and the lane it came from.
2 lanes
One barrier between them, enforced — not advisory. The clean lane is all that productises.
0
Unprovenanced fields the analyst will assert. It abstains before it guesses.
Design partners [[ logo ]] [[ logo ]] [[ logo ]] — cohort forming; names on signature, never before.
What you point it at

Enrich an entity to decide on it.

The same waterfall enrichment and agentic research, aimed at a world where the answer must hold up.

Find the deals worth the diligence — before anyone else.

Source real-asset and private-market opportunities by signal, not by list. Each candidate arrives pre-scored on readiness, with the evidence behind the score attached, so the desk spends its hours on the deals that survive a second look.

You askSurface owners of B-grade industrial assets in [[ corridor ]] showing refinance or succession pressure.
It returnsA ranked shortlist — each entity with the originating signal, its source, a freshness stamp, and a confidence band. Nothing asserted it can't cite.

Diligence playbooks, run by an analyst that cites.

Codify the desk's diligence as a reusable playbook. The analyst mines public record, filings, and licensed sources for each line item — and where it can't find ground truth, it says so, instead of filling the gap with a plausible guess.

You askBuild the ownership and encumbrance picture for this asset and flag anything inconsistent with the vendor's representation.
It returnsA dossier where every claim links to its source, conflicts are surfaced not smoothed, and confidence is calibrated per field.

KYC & AML files an auditor can follow.

Beneficial-ownership chains, PEP and sanctions context, source-of-funds signals — each assembled with the lineage a regulator expects. The barrier status of every input is explicit, so the file is defensible the day it's questioned, not reconstructed under pressure.

You askResolve the beneficial owners behind this entity and confirm the registered proprietor matches.
It returnsThe chain, each hop sourced and lane-tagged, with mismatches raised explicitly and an abstention where the record is silent.

The dossier stays alive after close.

The living-system thesis, pointed at risk: event-triggered monitoring across the book. A covenant signal, an ownership change, a freshness decay on a key field — the dossier updates itself and tells you which positions just changed, with the evidence.

You askWatch this portfolio for covenant-breach and refinance-risk signals; tell me what moved this week.
It returnsA change-log of positions, each delta carrying the triggering event, its source, and a freshness stamp — not a static snapshot.

Know the counterparty before you're exposed to it.

Profile a sponsor, borrower, or counterparty across record, litigation, and reputational signal — with every input's provenance and barrier status intact, so the risk view is one you can put in front of an investment committee without a caveat about where it came from.

You askAssemble a counterparty risk view on this sponsor across their entity network.
It returnsA network map and risk read, each node sourced and confidence-banded, with the clean lane clearly separated from anything that can't cross the barrier.
The showpiece

What do you need to prove?

Ask in plain language. What comes back isn't an answer — it's a defensible one.

Who controls 14 Marine Parade — and is the beneficial owner the registered proprietor?
Trace ownership chain Covenant signals KYC dossier Verify a representation
Returned datum — beneficial ownerClean lane ✓
Held via [[ Holdco Pty Ltd ]] — beneficial owner differs from registered proprietor.
Source
ASIC + Land Registry
Confidence
0.91
Freshness
retrieved 2d ago
Barrier
licensed · cleared

One field carries source, confidence, freshness, and barrier status. Multiply by the dossier — that's the audit trail, generated, not assembled.

Illustrative. Figures bracketed [[ ]] are placeholders — the structure is the point.

The platform

Four primitives. Each one, provenance-first.

01 · Provenance

One contract, every source — each value still tagged.

Buy from many data vendors under a single agreement, the way a marketplace does. The difference: nothing loses its lineage on the way in. Each value keeps its source, confidence, freshness, cost, and lane — permanently.

Where Clay has a data marketplace · here it is a chain of custody
02 · The Analyst

An analyst that does citable diligence.

The agentic researcher, recast as a junior analyst with discipline: it mines the web and licensed record for the data points a playbook needs, calibrates its confidence honestly, and abstains rather than fabricate. Its output is meant to be defended.

Where Clay has a research agent · here it is a diligence analyst
03 · The Barrier

Information barriers, enforced by construction.

The clean/restricted lane split isn't a setting — it's the architecture. The barrier finance is legally required to maintain is the same firewall the engine runs on. What can't cross, structurally cannot. This is the moat.

Where Clay has orchestration · here it is a governed barrier
04 · Living Dossier

Diligence that monitors itself after close.

Recipes become diligence playbooks; the living-system loop becomes event-triggered deal and risk monitoring. The dossier doesn't go stale on a shelf — it watches its own inputs and tells you when a decision's basis has changed.

Where Clay has play execution · here it is continuous monitoring
How the analyst works

From a question to a defensible answer.

01
Frame
Turn the question into the data points a playbook requires.
02
Source
Waterfall across licensed vendors and public record; tag each hit.
03
Cross-check
Corroborate or surface conflict — never silently reconcile.
04
Calibrate
Assign honest confidence; abstain where the record is silent.
05
Tag lane
Mark barrier status — clean stays clean, restricted stays out.
06
Return
A dossier where every claim carries its own evidence.
The boundary

What we will never do.

In a market built on trust you can't re-earn, the discipline that protects you is the same one that differentiates you. So we state the line plainly — and we sell only what's on the right side of it.

The product

The clean lane, and nothing but.

  • Licensed and first-party sources, each carrying its right-to-use.
  • Provenance visible on every datum — auditable the day it's questioned.
  • Calibrated confidence, with abstention as a first-class answer.
  • Information barriers enforced structurally, not by policy memo.
Off the table

What stays out of the productised offering.

  • Scraped or unlicensed data dressed up as enrichment.
  • Assertions without a source. If it can't be cited, it isn't returned.
  • Anything that lets restricted data cross the barrier.
  • Coverage claims we can't defend in front of a regulator.

The firewall stopped being an architecture decision. It became a brand boundary.

The climb

Beachhead to clearinghouse.

Real estate is a capital-markets asset class with retail flow — which is why we start where we already hold the data, the relationships, and the SMSF edge.

Today

Property & SMSF intelligence

The beachhead. We own the data moat, the Cotality / RP Data backbone, and the engine. Provenance pays for itself here first.

Next

Real-asset & private-market diligence

The same provenance-grade layer, generalised to the deals and counterparties where rigour is mandatory and paid for.

Endgame

The exchange

Provenance-carrying entities clear once, to one accountable buyer. The intelligence layer becomes the rail beneath it.

"Clay for capital markets" is the door. Provenance-native is the room.

Request access
The space between a sales tool and a clearinghouse is the distance from coverage to proof.

We built for proof. If you run an origination, diligence, or risk desk in real assets, let's put a defensible dossier in front of you.