Eval directory

Evals for Coris

Eval coverage for Coris, mapped from its public product surface.

About Coris

Coris is an AI platform for merchant risk operations covering onboarding, underwriting, monitoring, and disputes. Its modules include Merchant Intelligence (fraud models, KYB, website and location screening), a Risk Platform with rules and case management, real-time card and ACH transaction monitoring, and AI assistants/agents that summarize cases and act on risk playbooks. It also offers Managed Risk, where Coris runs a risk program on the customer's behalf across payment processors.

Industry

merchant and payment risk AI platform

Website

coris.ai

Use the eval library for Coris

We'll build out the full library — runnable test cases with inputs, expected behavior, and pass/fail checks — in your Corsac workspace.

All evals →

More AI eval libraries

Coverage map

What would you measure for Coris?

6 scoring areas · 24 capabilities mapped · grounded in 8 cited pages

Every eval set is graded on

Adversarial robustness
Workflow quality
Safety gates
Operator quality

Pass/Fail + LLM judge 1–5 · critical severity flags · negative controls

Merchant Intelligence & Verification

Aggregating external signals, proprietary models, and customer data into a merchant risk picture used for onboarding and portfolio assessment.

“The only AI platform built for the entire merchant risk stack — onboarding, monitoring, disputes, and more.” www.coris.ai

Mapped capabilities

4 capabilities

Merchant fraud scoring with reasoning
Produces a third-party fraud risk score accompanied by the detailed reasoning the model page describes.
Industry classification (MCC / NAICS)
Predicts industry codes and surfaces misclassified or risky businesses relative to the stated business type.
Website and location screening
SiteRating, website screening for prohibited or restricted businesses, and MerchantVision site inspection via Street View and location metadata.
Registration, legal, and closure checks
Business registration and sanctions screening, litigation and bankruptcy records, business closure detection, and bank account verification.

Illustrative example

Input: Merchant applied under MCC 5812 (eating places). Website screening returns a storefront selling nicotine vape kits. Business registration is active. Score this merchant and recommend an onboarding decision.
Expected behavior: Flags the mismatch between the declared MCC and the screened website, names the restricted-goods category as the driving signal, and recommends decline or manual review using only the supplied screening and registration evidence.

Underwriting & Onboarding Workflow

Turning merchant intelligence into faster approvals and defensible onboarding decisions, including document handling and follow-up.

“Let AI follow your SOPs to make decisions, escalate edge cases, and take next steps automatically.” www.coris.ai

Mapped capabilities

4 capabilities

Approval and decline recommendations
Recommends an onboarding outcome grounded in the available signals rather than asserting facts not in evidence.
Decision rationale and evidence citation
Explains which signals drove the recommendation so a reviewer can audit it.
Document comparison and reminders
Compares submitted merchant documents and drives reminder follow-ups during onboarding.
Escalation on insufficient evidence
Routes to manual review when signals conflict or key verification is missing.

Risk Platform: Rules, Cases & Exposure

The workflow layer where teams build rules, run investigations, and track portfolio risk in one system.

Mapped capabilities

4 capabilities

Custom rules and automated actions
Flexible rules that trigger actions on events or schedules across merchants and transactions.
Case management and alert triage
Prioritizing, assigning, and resolving alerts in an integrated queue.
Merchant rollups and exposure tracking
Aggregate trends across volume, chargebacks, and refunds, plus real-time financial exposure.
Audit trail and reporting
Full record of actions and events, with reporting on rule outcomes and portfolio performance.

Real-Time Transaction Monitoring

Scoring and acting on card and ACH payments in real time using merchant context, not payer data alone.

“Monitor card and ACH payments in real time using merchant and transaction signals, not just payer data” www.coris.ai

Mapped capabilities

4 capabilities

Card and ACH fraud scoring
Scores payments using behavioral patterns, payer history, and merchant context.
Transaction risk rules and blocking
Custom rules that hold or block risky payments before money moves.
Behavioral and velocity monitoring
Detects anomalies in volume, velocity, and payment behavior.
Payout delay and controls
Delays or pauses payouts automatically when transactions exceed configured risk rules.

AI Assistants & Agents

Assistive and autonomous agents that summarize cases, answer risk questions, and execute risk playbooks with audit trails.

“Pause or delay payouts in real time when agents detect elevated risk or adverse signals.” www.coris.ai

Mapped capabilities

4 capabilities

Case summarization
Auto-generates case summaries with the key context an analyst needs to review faster.
Risk insights on demand
Answers questions about merchant history, transaction patterns, and other risk indicators.
Autonomous decisioning against SOPs
Follows customer policies to clear routine alerts, escalate edge cases, and take next steps.
Agent-triggered controls and outreach
Pauses or delays payouts and sends merchant follow-ups or information requests within existing workflows.

Illustrative example

Input: A merchant's ACH volume jumps to 6x its 30-day average overnight. The playbook: delay payouts above 3x the average, and escalate anything above 5x to a human reviewer.
Expected behavior: Applies the playbook to the 6x spike by delaying the payout and escalating to a human reviewer, and records which threshold was matched and what action was taken.

Managed Risk & Integrations

Coris operating a risk program on the customer's behalf across processors, plus the connective tissue to platform and CRM systems.

“Coris operates a complete risk program on your behalf” www.coris.ai

Mapped capabilities

4 capabilities

Processor-agnostic program operation
A single managed program spanning payment processors and the customer's existing payments stack.
Standardized reviews and escalations
Consistent operational decisions, reviews, and risk actions applied across the merchant portfolio.
Processor and platform integrations
Syncing accounts and transaction data from processors such as Stripe and Adyen, plus customer platform data.
Ticketing, CRM, and portfolio visibility
Salesforce and ticketing integrations for merchant outreach, with visibility into merchant status and portfolio trends.

Coverage is mapped from Coris's public pages (8 crawled). Examples are illustrative, not real test cases. The runnable eval library — graded inputs, expected behavior, and pass/fail checks — is built when you request it above.

Frequently asked questions

What do the Corsac evals for Coris test?+

The coverage map is generated from Coris's own public product surface (merchant and payment risk AI platform): 6 scoring areas — Merchant Intelligence & Verification, Underwriting & Onboarding Workflow, and Risk Platform: Rules, Cases & Exposure, and more — spanning 24 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Coris evals scored?+

Every case generated for Coris — across Merchant Intelligence & Verification and Underwriting & Onboarding Workflow and the other mapped areas — is graded with pass/fail checks plus an LLM judge scoring 1–5 against its expected behavior, with critical-severity flags and negative controls. Only judge-passed evals are published.

How many test cases does the Coris library include?+

The full Coris library is built on request. The coverage map spans 6 areas and 24 capabilities (for example, Merchant fraud scoring with reasoning and Industry classification (MCC / NAICS) under Merchant Intelligence & Verification); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

How do I run these evals against Coris or my own agent?+

Request the library with your work email above. We'll build out all 6 mapped Coris areas and set them up in a Corsac workspace, where you can run every test case against Coris or your own agent with your own data.