What is an Account Confidence Score?

An Account Confidence Score is a composite risk rating for a counterparty — vendor, customer, or partner — that aggregates multiple risk signals (OFAC sanctions status, fraud indicators, behavioral anomalies, duplicate payment risk) into a single 0–100 score, where higher means lower risk.

What data sources are used to calculate the score?

The score combines: OFAC SDN and consolidated sanctions list checks, vendor master deduplication analysis, payment behavior patterns, TIN/EIN validation, address consistency checks, and configurable custom risk rules.

How is the Account Confidence Score API different from standalone OFAC screening?

OFAC screening is a binary check: is this name on a sanctions list? The Account Confidence Score is multi-dimensional — it layers sanctions risk with fraud signals, behavioral risk, and structural anomalies into a single composite metric with an explanation.

What score threshold should I use for automatic payment blocking?

Typical configurations use: score 80–100 (auto-approve), score 50–79 (manual review), score 0–49 (block and escalate). Your optimal thresholds depend on your risk tolerance and industry. The API returns the component breakdown so you can calibrate thresholds for your use case.

Can the score be integrated into an ERP or AP automation workflow?

Yes. The Account Confidence Score is available as an MCP tool (through Claude) and as a REST API endpoint that can be called from any ERP, AP automation platform, or custom workflow. See the API docs for integration details.

Is this compliant with BSA/AML requirements?

The score includes OFAC screening which is a component of AML compliance. However, ClaudeFinLab's tools are analytical aids, not certified compliance solutions. Your compliance program should be designed and reviewed by qualified counsel. The tool helps automate screening that your compliance team then reviews.

What is the Bad Debt Provision Estimator and how does it relate to account scoring?

The Bad Debt Provision Estimator estimates expected credit losses on receivables portfolios, often using account-level risk scores as inputs. Running the Account Confidence Score first, then passing high-risk accounts to the Bad Debt Provision Estimator, creates an end-to-end counterparty risk workflow.

How fast does the scoring API respond?

Single account scoring typically returns in under 2 seconds. Batch scoring of 100+ accounts runs in under 30 seconds. Real-time integration (e.g. pre-payment check) is feasible.

What is duplicate payment detection?

Duplicate payment detection identifies when the same invoice is being submitted twice — often with slight variations in vendor name, invoice number, or amount. It's a common fraud vector in AP and a key component of the Account Confidence Score.

Does the score work for individual accounts or only business entities?

The score works for both business entities (vendors, counterparties) and individuals (contractors, employees in expense workflows). Individual scoring includes PEP (Politically Exposed Person) checks and personal OFAC screening.

What is the difference between counterparty risk and credit risk?

Credit risk is the probability that a counterparty fails to repay a debt or obligation. Counterparty risk is broader — it includes credit risk, but also operational risk (is this entity real?), compliance risk (is this entity sanctioned?), and fraud risk (is this a shell entity or duplicate vendor?). The Account Confidence Score covers all three dimensions.

Can I customise which risk factors are weighted in the score?

Yes. Enterprise users can configure component weights via the API — for example, a bank might weight sanctions risk heavily while a mid-market company might weight duplicate vendor risk more highly. Default weights are pre-configured for general-purpose use.

How does the score handle partial name matches in sanctions screening?

The sanctions component uses fuzzy matching with a configurable similarity threshold (default: 85%). Matches above the threshold generate a flag; Claude explains the similarity and the specific SDN entry that matched so a human reviewer can make the final call.

What is the pricing for the Account Confidence Score API?

The score is available free at 100 calls/day through ClaudeFinLab's standard free tier. Higher-volume enterprise plans are available on request. Each account score counts as one API call.

How do I audit which risk factors caused a low score?

Every score comes with a component breakdown: OFAC score, fraud signal score, behavioral score, structural score, and a plain-English explanation of each factor that contributed to the composite rating. This makes the score auditable and defensible to regulators.

Is this tool useful for KYC workflows?

The Account Confidence Score handles the screening and risk-assessment layer of KYC, but it is not a KYC platform. Full KYC requires document verification and identity proofing steps that are outside ClaudeFinLab's scope. However, for the risk-scoring component of KYC, the API provides significant automation.

Can I screen a list of existing vendors for retroactive compliance?

Yes. Batch scoring via the API is designed for exactly this use case — upload your vendor master file and receive a scored, ranked list of vendors by risk level, with the highest-risk flagged for immediate review.

What is the vendor master file fraud problem?

Vendor master fraud involves adding fictitious or duplicate vendors to an organization's AP system to divert payments. Common patterns include ghost vendors (fake entities), duplicate vendors (same real entity added twice with slight name variations), and insider fraud (employee-controlled vendors). The Account Confidence Score flags all three patterns.

How does the score integrate with the OFAC screening skill?

The Account Confidence Score includes OFAC screening as one component. If you only need OFAC checks (not the full composite score), the standalone Vendor Master File Sanction & Fraud Screener skill is available and covers OFAC, EU, and UN sanctions lists.

Can I use this for vendor onboarding workflows?

Yes — vendor onboarding is the primary use case. Before adding a new vendor to your ERP and making a first payment, run the Account Confidence Score. It takes seconds and surfaces OFAC flags, duplicate vendor risk, and structural red flags at the point of onboarding.

What happens when a vendor is flagged?

The tool generates an explanation of the specific flag — for example, 'Name is 91% similar to SDN entry ACME TRADING LLC (ID: 12345)' or 'Bank routing number matches 3 other vendors in your master file — possible duplicate.' Your compliance or AP team then reviews and makes the final call.

How does this help with SOX compliance?

SOX 404 requires documented controls over financial reporting, including AP processes. Using an Account Confidence Score to screen vendors before payment creates an auditable control with consistent application — supporting your SOX 404 testing and demonstrating effective controls to auditors.

Is there a webhook for real-time payment approval flows?

Webhook integration is available on enterprise plans, allowing the score to trigger in real-time when a payment is submitted in your ERP, blocking low-confidence payments before they are processed.

Account Confidence Score API: AI-Powered Counterparty Risk Screening

ClaudeFinLab's Account Confidence Score API flags high-risk counterparties before payment. Screen vendors, clients, and accounts against fraud signals, OFAC sanctions, and behavioral risk factors - through Claude in seconds.

What is the Account Confidence Score?

The Account Confidence Score is a composite 0–100 risk rating for any counterparty — vendor, customer, contractor, or financial account — that aggregates multiple risk signals into a single actionable metric. A score of 100 means high confidence (low risk); a score below 50 means the account warrants immediate review before payment.

ClaudeFinLab exposes this as an MCP tool through Claude and as a REST API. Finance and compliance teams use it to automate vendor screening, flag suspicious payments before they process, and maintain a continuously scored vendor master file.

Why Account Screening Matters More Than Ever

Payment fraud, vendor master manipulation, and sanctions violations are not hypothetical risks. Finance teams deal with them regularly:

Ghost vendor fraud: fictitious suppliers added to the AP system to divert payments
Duplicate vendor fraud: real suppliers added twice with slight name variations
Sanctions violations: inadvertently paying an OFAC-designated entity — triggering civil or criminal liability
Invoice fraud: realistic-looking invoices from spoofed or newly registered entities
Business email compromise: supplier account detail changes submitted by attackers

Manual screening catches some of these. Automated scoring catches all of them consistently, at zero marginal cost per check.

Score Components

Component	Weight	What it checks
OFAC/Sanctions	35%	SDN, consolidated list, fuzzy name matching
Structural Validity	25%	EIN/TIN format, address consistency, entity age
Duplicate Detection	20%	Name similarity to existing vendors, bank account overlap
Behavioral Signals	20%	Payment pattern anomalies, invoice rounding, frequency changes

Configuration: Enterprise users can adjust component weights to match their risk tolerance. A bank might weight OFAC at 50%; a manufacturing company might weight duplicate detection higher.

Score Interpretation

Score range	Risk level	Recommended action
80–100	Low risk	Auto-approve payment / onboard vendor
60–79	Moderate risk	One-level approval required
40–59	Elevated risk	Compliance team review
0–39	High risk	Block and escalate immediately

Using the API Through Claude

Get your free API key at claudefinancelab.com
Add the compliance MCP server to Claude Desktop:

{
  "mcpServers": {
    "claudefinlab-compliance": {
      "url": "https://claudefinancelab.com/compliance/sse",
      "headers": { "Authorization": "Bearer YOUR_API_KEY" }
    }
  }
}

Ask Claude: "Score account confidence for vendor: Acme Supplies LLC, EIN 12-3456789, address 100 Main St, Dallas TX. Flag if OFAC match or duplicate of existing vendors."

Batch Vendor Master Screening

For retroactive compliance, the batch API allows you to score your entire vendor master in one call:

POST /compliance/account-confidence-batch
{
  "accounts": [
    {"name": "Acme Supplies LLC", "ein": "12-3456789", "address": "..."},
    {"name": "Global Trade Partners", "ein": "98-7654321", "address": "..."},
    ...
  ],
  "flag_threshold": 60
}

The response includes each account's composite score, component breakdown, and specific flags — ranked from highest to lowest risk for efficient review prioritization.

Related Compliance Tools

Vendor Master Sanction & Fraud Screener — dedicated OFAC + fraud screening for vendor onboarding
P2P Duplicate Payment Preventer — catch duplicate invoices before payment
SOX 404 Controls Testing — demonstrate effective AP controls to auditors
Bad Debt Provision Estimator — estimate expected losses on high-risk receivables

Get your free API key

Account Confidence Score API: AI-Powered Counterparty Risk Screening

What is the Account Confidence Score?

Why Account Screening Matters More Than Ever

Score Components

Score Interpretation

Using the API Through Claude

Batch Vendor Master Screening

Related Compliance Tools