Recipes
Ready-to-run Grayson requests for fraud, payments, disputes, onboarding and AML decisions, with real responses and the cost and latency of each decision.
Each recipe is one decision a fraud, risk or compliance team makes every day, written as a Grayson request: the data to send, the questions to ask, a real response, and how to add your own policy. Every recipe ends with the steps to evaluate it on your own data.
Cost is for each recipe's example request at $0.035 per million input tokens. Latency is the median of several calls of that request through api.finic.ai from US-West; add your own network time from elsewhere.
Fraud
| Recipe | Decides | Cost per 1,000 decisions | Latency (US-West) |
|---|---|---|---|
| Detect insider threats in employee account access | Decide whether an employee's account access is work, snooping or insider fraud, and close, refer or suspend. | $0.06 | 181 ms |
| Fraud typology classification | Label a fraud claim with a FraudClassifier class, scam type and contributing factors; decide reimbursement. | $0.13 | 214 ms |
| Triage dark web credential and card alerts | Match a dark web alert to the customer, judge whether the exposure still works, and choose the response. | $0.07 | 184 ms |
Payments
| Recipe | Decides | Cost per 1,000 decisions | Latency (US-West) |
|---|---|---|---|
| Detect business email compromise in vendor payments | Catch payments to a vendor's new bank account after a fake or compromised change request. | $0.05 | 184 ms |
| Detect check fraud in mobile and branch deposits | Release, hold or reject a check deposit that may be counterfeit, altered, stolen or a duplicate. | $0.04 | 172 ms |
| Detect fraudulent incoming ACH credits and transfers | Release, hold or return an incoming credit that may be proceeds of BEC, account takeover or payroll diversion. | $0.05 | 173 ms |
| Screen card authorizations for fraud in real time | Score card authorizations for fraud and choose approve, step up or decline, next to your existing rules. | $0.04 | 168 ms |
Disputes
| Recipe | Decides | Cost per 1,000 decisions | Latency (US-West) |
|---|---|---|---|
| Classify card disputes and detect friendly fraud | Classify unauthorized-transaction claims and choose provisional credit, approval or denial. | $0.06 | 219 ms |
Onboarding
| Recipe | Decides | Cost per 1,000 decisions | Latency (US-West) |
|---|---|---|---|
| Detect account opening fraud and synthetic identities | Approve, step up or decline a deposit account application at risk of stolen or synthetic identity fraud. | $0.05 | 183 ms |
AML
| Recipe | Decides | Cost per 1,000 decisions | Latency (US-West) |
|---|---|---|---|
| Detect money mule accounts | Whether an account is passing other people's money through, whether the holder knows, and what to do | $0.08 | 208 ms |
Evaluate on your own data
Every recipe publishes its questions at /recipes/<recipe>/questions.json, so you can score Grayson on your own cases with one command. See Evaluate on your own data.
Evaluate on your own data
Score Grayson on your own past cases with the grayson CLI. Point it at a CSV, ask your questions, and see how often Grayson gives the answer you know is right.
Account opening fraud
Grayson approves, steps up or declines a deposit account application, weighing identity-vendor results, credit file, contact, device and velocity data.