CloudSketch AI Logo
FeaturesTemplatesPricingEnterpriseAboutContactLog in
πŸŒ™β˜€οΈ
Log inStart free
Home/Templates/Finance & Banking

🏦 Finance & Banking · Data Flow

Real-Time Fraud Detection Data Flow

How a bank scores every transaction for fraud in milliseconds: features from history, an ML model, rules, and a case queue for analysts.

More Finance & Banking templates

Opening a template is free and never counts toward your daily diagrams. Asking the AI to change it counts as one.

Drawing diagram…

What this diagram shows

  • Each transaction is scored before it is approved
  • Features combine the live transaction with recent history
  • Borderline cases go to analysts; their decisions retrain the model

Prompt used

Data flow for real-time fraud detection: card and UPI transactions stream through Kafka to a scoring service, which pulls features (spend velocity, device, location, merchant risk) from a feature store, runs an ML model and business rules, and returns approve, decline or step-up authentication to the payment switch. High-risk cases go to an analyst case queue; analyst decisions are labelled and feed model retraining. Number the processes and label every flow.

Mermaid code
flowchart LR
subgraph EXT ["External entities"]
payment_switch["Payment Switch"]
fraud_analyst["Fraud Analyst"]
end
subgraph PROC ["Processes"]
p1("1.0 Ingest transaction stream")
p2("2.0 Fetch transaction features")
p3("3.0 Evaluate fraud risk")
p4("4.0 Route high-risk cases")
p5("5.0 Review and label cases")
p6("6.0 Retrain ML model")
end
subgraph STORE ["Data stores"]
d1[("D1 Kafka Transaction Stream")]
d2[("D2 Feature Store")]
d3[("D3 Analyst Case Queue")]
d4[("D4 Labelled Fraud Dataset")]
d5[("D5 ML Model Registry")]
end
payment_switch -- "Raw transaction stream data" --> p1
p1 -- "Streamed transaction data" --> d1
d1 -- "Real-time transaction event" --> p2
d2 -- "Spend velocity, device, location, and merchant risk features" --> p2
p2 -- "Enriched feature vector" --> p3
d5 -- "Trained ML model parameters" --> p3
p3 -- "Transaction decision (Approve, Decline, Step-Up)" --> payment_switch
p3 -- "High-risk transaction score and details" --> p4
p4 -- "Flagged case record" --> d3
d3 -- "Pending case data" --> p5
p5 -- "Analyst decision and investigation label" --> d4
p5 -- "Manual review outcome" --> fraud_analyst
fraud_analyst -- "Case disposition and analyst label" --> p5
d4 -- "Historical labelled transaction data" --> p6
p6 -- "Updated ML model artifact" --> d5

Inside this diagram

A data flow diagram with 8 processes and outside parties, 5 data stores and 15 data flows.

Processes and outside parties

  • Payment Switch
  • Fraud Analyst
  • 1.0 Ingest transaction stream
  • 2.0 Fetch transaction features
  • 3.0 Evaluate fraud risk
  • 4.0 Route high-risk cases
  • 5.0 Review and label cases
  • 6.0 Retrain ML model

Data stores

  • D1 Kafka Transaction Stream
  • D2 Feature Store
  • D3 Analyst Case Queue
  • D4 Labelled Fraud Dataset
  • D5 ML Model Registry

Data flows

  • Payment Switch β†’ 1.0 Ingest transaction stream: Raw transaction stream data
  • 1.0 Ingest transaction stream β†’ D1 Kafka Transaction Stream: Streamed transaction data
  • D1 Kafka Transaction Stream β†’ 2.0 Fetch transaction features: Real-time transaction event
  • D2 Feature Store β†’ 2.0 Fetch transaction features: Spend velocity, device, location, and merchant risk features
  • 2.0 Fetch transaction features β†’ 3.0 Evaluate fraud risk: Enriched feature vector
  • D5 ML Model Registry β†’ 3.0 Evaluate fraud risk: Trained ML model parameters
  • 3.0 Evaluate fraud risk β†’ Payment Switch: Transaction decision (Approve, Decline, Step-Up)
  • 3.0 Evaluate fraud risk β†’ 4.0 Route high-risk cases: High-risk transaction score and details
  • 4.0 Route high-risk cases β†’ D3 Analyst Case Queue: Flagged case record
  • D3 Analyst Case Queue β†’ 5.0 Review and label cases: Pending case data
  • 5.0 Review and label cases β†’ D4 Labelled Fraud Dataset: Analyst decision and investigation label
  • 5.0 Review and label cases β†’ Fraud Analyst: Manual review outcome
  • Fraud Analyst β†’ 5.0 Review and label cases: Case disposition and analyst label
  • D4 Labelled Fraud Dataset β†’ 6.0 Retrain ML model: Historical labelled transaction data
  • 6.0 Retrain ML model β†’ D5 ML Model Registry: Updated ML model artifact

Boundaries

  • External entities: Payment Switch and Fraud Analyst
  • Processes: 1.0 Ingest transaction stream, 2.0 Fetch transaction features, 3.0 Evaluate fraud risk, 4.0 Route high-risk cases, 5.0 Review and label cases and 6.0 Retrain ML model
  • Data stores: D1 Kafka Transaction Stream, D2 Feature Store, D3 Analyst Case Queue, D4 Labelled Fraud Dataset and D5 ML Model Registry

Make your own diagram like this

Describe your own system in plain English and CloudSketch AI draws it in seconds. See all AI diagram generators.

Related templates

SequenceFinance & Banking

UPI Payment Flow

The steps behind a UPI payment: the payer's app, the payer's bank, NPCI's UPI switch and the payee's bank, from entering a UPI ID to the money arriving.

C4 ArchitectureProFinance & Banking

Payment Gateway Architecture

A payment gateway that lets online merchants accept cards and UPI: the checkout, the payment core with fraud checks, and the card networks and banks it connects to.

FlowchartFinance & Banking

Loan Origination Workflow

A personal loan from application to disbursal: KYC, credit bureau check, underwriting rules, manual review for borderline cases, and the final agreement.

SequenceFinance & Banking

Card Payment Authorisation

What happens in the second after a card is tapped or entered online: the acquirer, card network and issuing bank authorise the payment, with 3-D Secure for online payments.

C4 ArchitectureProFinance & Banking

Core Banking System Architecture

The main building blocks of a bank's core platform: accounts, deposits, loans, payments and the general ledger, used by branches, internet banking and mobile.

FlowchartFinance & Banking

Digital KYC Onboarding

How a bank or fintech opens an account fully online: Aadhaar/PAN verification, video KYC, risk checks and account creation.