AnantState
For the Chief Risk Officer and fraud leads

Work the case that matters first.

Cases are ranked by what each costs if ignored, with the evidence behind them and the places where sources disagree. Holds, reviews and step-up checks each carry a price, so the false-positive trade-off becomes a decision someone can review.

In plain termsCases ranked by money at stake, with the evidence and where it disagrees.

At a glance

  1. Ranked by money at stake

    The queue is ordered by value, not by how unusual a case looks.

  2. Conflict kept in view

    Where sources disagree the case is marked contentious, not averaged away.

  3. Priced actions

    A hold, a review or a step-up check, each with its cost.

  4. A defensible record

    Evidence, decision, actor and time on every case.

The questions on a risk leader’s desk

  • Which of thousands of alerts matters?
  • What are our false positives costing, and who is pricing them?
  • Where do our sources disagree, and does anyone see it?
  • Could we defend this decision to a regulator?

One case, start to finish

A case is ranked with its evidence, an analyst approves a priced action, and the evidence, decision, actor and time go on the audit trail.

One case, start to finish

  1. 1. Account activity to AnantState: Behavior breaks from its history
  2. 2. AnantState to An analyst: A ranked case, sources in conflict
  3. 3. An analyst to AnantState: Approves a priced hold
  4. 4. AnantState to Audit trail: Evidence, decision, actor, timea reply or asynchronous
A case is ranked with its evidence, including where sources disagree, an analyst approves a priced action, and the evidence, decision, actor and time go on the audit trail.

Why the disagreement matters

A weighted score averages two weak sources and two flatly contradicting ones to the same number. Evidence fusion keeps the contradiction, because a contested case is the one an experienced analyst should look at.

Same average, different case · illustrative

Same average, different case · illustrative, illustrative values
CategorySource ASource BResult
Weak agreement0.50.5Low conflict
Contradiction10High conflict
A weighted score averages these two cases to the same number. Evidence fusion keeps the disagreement, because a contentious case is the one a person should look at.

Concerns a risk leader is likely to raise

The concernThe answer
Model risk will want to validate it.The model record is built for that: documented purpose, performance against a defined baseline, a recorded promotion decision and drift monitoring.
We can’t let a system act on contested evidence.It doesn’t. The conflict gate holds the action for a person regardless of the autonomy level.
Does it claim regulatory compliance?No. We state the controls that exist and your compliance function maps them to the regulation.

Bring a queue

Bring:

  • A queue
  • The rule nobody has changed in two years

You leave with the queue re-ordered by value at stake, and that rule priced against the loss it prevents.

Price the false positive

Bring a rule whose false-positive cost nobody has measured. We will price it against the cost of the loss it prevents.