Talk to us

Use cases

One engine for everything that should look normal.

Proven on financial crime. Built to learn what normal looks like anywhere.

Financial crime

Benchmarked

Laundering hides in the shape of money movement.

Rings and pass-through accounts only show up across the network and over time. That is where the agent looks.

Runs onTransfers, accounts, customers, cash activity

Scored on three public benchmarks. See the results

Ask it

Flag accounts that take in many small transfers from new counterparties and move most of it out within a day.

Flag groups of accounts that route funds in a loop back to where they started.

Security

Pack in development

Intrusions are anomalies with intent behind them.

Stolen credentials look normal until you compare the user, host and service with their usual selves.

Runs onAuthentication logs, network flows, endpoint events, cloud audit trails

Looking for design partners. Talk to us

Ask it

Flag service accounts that sign in from a host they have never used, outside their usual hours.

Flag hosts whose outbound volume to new destinations jumps against their own baseline.

Fraud

Same engine

Fraud moves faster than a rule backlog.

Takeover and mule patterns change every quarter. A detector described in a sentence keeps pace.

Runs onLogins, devices, payments, payouts, disputes

Same engine; not yet benchmarked. Bring a dataset

Ask it

Flag new accounts that share a device or payout details with accounts closed for fraud.

Flag a password reset followed within the hour by a new payee and a large transfer.

Operations

Tested

Machines drift before they fail.

The same loop, pointed at sensors. Each machine is compared with its peers.

Runs onSensor readings, maintenance logs, quality checks

Tested on public predictive-maintenance data. How it generalises

Ask it

Flag machines whose torque and tool wear drift away from peers of the same type.

Flag lines where the reject rate rises before any single sensor looks wrong.

Something else

Claims, trading, health, gaming, supply chains.

If your data has a normal, the engine can learn it.

Talk to us