Capture the bounded workflow, environment and desired outcome.
EcoKure demos · one place
See one decision. Then see what survives change.
Start with the plain transaction example. Then open the Control Tower, change a rule or source, and inspect the evidence path. Everything below is synthetic or illustrative unless marked as a target-environment measurement.
Demonstration boundary: no customer data, no production credentials and no live workflow actuation are used here.
01 · Existing workflow → governed decision
Invented names and amounts“Show me this one transaction.”
Pick your industry, then drag one of its cases onto the board. Each one goes through two rules your side owns. Use the out-of-date list on purpose and watch it refuse to treat a stale result as safe.
Pick your industry
Which list gets checked
Cases — drag one onto the board, or press one
Drop a case here
Or press one of the cases on the left.
What this proves — and what it does not
It proves: the example evaluates against versioned rules and produces a signed record. It does not prove: that EcoKure is an AML system, decides suspicion, files anything, provides legal advice or certifies a casino’s compliance.
02 · Product view
Illustrative dataControl Tower: the operating picture
This view explains what a buyer sees after the decision is inside the control boundary: recent outcomes, named review, change impact and evidence.
Workspace / enterprise-ai
illustrative streamDecision stream
Systems12monitored
Control packs38active
Review queue2needs a person
Evidence100%chain complete
Recent governed decisions
What the tower is seeing
Select a decisionIts control result and evidence reason will appear here.
Named human action
2 OPENReview queue
Claims triage assistant
REVIEWHigh-impact recommendation · owner: Risk Operations
Clinical literature assistant
REVIEWIncomplete provenance · owner: Clinical Quality
Review record readyIn a customer deployment, the decision, reason, identity and resolution would be appended to the evidence chain.
Control lineage
REPLAY REQUIREDChange impact centre
→→→
Select a nodeThe graph follows the evidence dependency, not only the software topology.
Independent verification
CHAIN HEALTHYEvidence register
EK-04291 · signed decision record
REPLAYABLEInput digest · control pack v3.2 · verdict · reviewer state · timestamp
sha256: 4d7a…9c12EvidencePack / 2026-08-27
COMPLETEPolicy, control, decision and change lineage records
offline verifier compatibleEvidence survives the product boundaryExported records are designed to be checked during procurement, audit and exit review.
03 · Change, replay and proof
Synthetic BioChangeGateWatch a source change become a governed outcome
A source revision changes. EcoKure separates affected, preserved and unknown work, replays what is deterministic and holds the rest for qualified review.
2affected records identified
3preserved records not replayed
1unknown records held for review
Create a customer boundary with explicit ownership and scope.
Issue a tenant-scoped key once and transfer it securely.
Agree the workflow owner, source and acceptance criteria.
Version your rules on where data came from, who reviews it and what evidence is needed.
Observe classifications without actuating a customer workflow.
Introduce one known revision to the reference source.
Separate the work the change touched, and re-run only that.
Hold uncertainty for a named qualified reviewer.
Verify the signed chain and record the next decision.
Demo boundary. The chain is demo-signed, uses fixture data and does not represent a medical, clinical or regulatory decision. A customer pilot still requires customer-owned keys, source boundaries, acceptance criteria and human sign-off.
04 · Regulated workflow lab
Synthetic onlyTest the pattern against a regulated workflow
Choose an example such as AML/CTF, KYC or safer-gambling controls. This lab measures the control-and-evidence pattern; it does not make a legal or regulated decision.
No run yet.
Observed result
SYNTHETIC · NO CUSTOMER DATA—
THROUGHPUT—transactions/sec
P50—milliseconds
P95—milliseconds
P99—milliseconds
ERRORS—recorded
AWS target pathReference environment, not production approval
117,290completed verifications
19.47 msp95 · 2,000 requests
128.1/sseparate soak run
51.5 secrecovery · 0 lost
Ready to test one real workflow?
Bring one owner, one bounded workflow and one evidence question.