Best AI for CFOs

Research notes on AI in finance work, by task

Index
W/P 11
Built
4 November 2026
Volume
1
This page
3 exhibits
  1. Best AI for CFOs
  2. Use cases
  3. Payment anomalies and fraud

11of 12 workpapers

Anomaly and fraud detection in payments

Can AI products catch fraudulent or anomalous payments before money leaves, and what controls must stay with people?

Published . Vendor pages read 29 October 2026.

How AI products flag anomalous payments, changed bank details and suspicious payees, the controls that remain with people, and questions for a vendor.

Products listed
12
Makers in the SEC ticker file
1
Filing excerpts
2
Notes
27

Lead schedule.Data the makers in this workpaper name

One stroke per record, in fives. Systems named by 4 or more records; every system is in Appendix C.

The schedule scrolls sideways; the workpaper column stays in place.

W/PBanksNetSuiteERPSAPCardsIntacctWorkdayClaudeCRMQuickBooksDynamicsOraclePaymentsPayrollSalesforceNo data named
11Payment anomalies and fraud212211111116

Answer sheet.What the makers' pages in this workpaper answer

One stroke per record: solid where the maker's own page answers the question, dotted where the pages read are silent. A count of what the pages say, not a score.

W/PData it namesWhere a person reviewsModel or data handling
11Payment anomalies and fraud6 of 125 of 125 of 12

1The work

AI products in payment fraud look for payments, payees and changes that do not fit the pattern (a new bank account on a long-standing supplier, a duplicate invoice, an unusual amount or timing, a payee linked to a blocked one) and hold them for a person before money moves. The makers in this note describe anomaly and fraud checks inside AP and spend platforms, payee risk screening, risk prioritisation for auditors, and real-time payment screening for banks.

The threat is well documented. The FBI's Internet Crime Complaint Center publishes an annual report on reported internet crime, including business email compromise.1 FinCEN has issued an alert on fraud schemes that use deepfake media to target financial institutions.2 The common pattern in business payments is simple: someone persuades the company to pay a real-looking invoice to a fraudster's account, often by impersonating a supplier or an executive.

Detection works on data the finance team already has: the vendor master and its change history, invoice and payment history, approval records and bank details. Some checks are rules (an account changed in the last 30 days); others are models trained on many transactions. Supplier verification has public tools too; the IRS says TIN Matching "is only for payers and their authorized agents that submit information returns."3

A person must stay in the loop for the release of any held payment, for any change to bank details (verified by a call to a number already on file, not one in the request), and for the decision to report or pursue suspected fraud. A flag that nobody reviews is not a control.

2Where AI is applied

Makers describe anomaly and fraud detection on bills and invoices, including generated receipts; comparison of payment details against vendor records and history; payee risk checks across shared details; prioritisation of high-risk activity for auditors; and real-time screening of payments.

Exhibit 11.1.What the makers say, in their words, alphabetical by maker

  • Catch anomalies, suspicious bill variances, and potential fraud before money moves out the door.10
    BILL, AI in BILL
  • Prevent payment threats in real-time with Payments Fraud Defense, a cloud-based, highly extensive, and flexible risk mitigation platform.11
    Bottomline, Payments Fraud Defense
  • Reduce risk with AI-powered fraud detection and out-of-the-box compliance automation, save time with PO matching, and optimize cash flow.12
    Coupa, Coupa AI, Coupa Navi agents
  • Machine-learning models scan every payment file in real-time and adapt as fraud tactics evolve.14
    Eftsure, Eftsure
  • AI-powered forecasting that's accurate, explainable, and continuously learning.15
    Kyriba, Kyriba Liquidity Performance Platform
  • Anomaly and fraud detection, including AI-generated receipts16
    Medius, Medius AI, Medius Copilot
  • MindBridge analyzes 100% of the financial data in scope, surfaces unusual activity, and explains what drives risk17
    MindBridge, MindBridge, Autonomous Financial Oversight
  • Surface high-risk T&E activity first, including repeat misuse, suspicious patterns, duplicate reimbursements, fake receipts, and policy violations.18
    Oversight, Travel and expense monitoring
  • Agents run the full flow from purchase order to posted payment: 3-way invoice matching, duplicate payments caught across entities and payment methods, contract rates enforced on every vendor transaction, and payment terms validated against what was negotiated.19
    Safebooks AI, Safebooks AI agents
  • Billy learns your coding patterns and applies them across your ERP fields.20
    Stampli, Billy, Stampli AI
  • Tipalti Accounts Payable embeds AI from invoice capture to ERP sync and reconciliation.21
    Tipalti, Tipalti AI, AP Agent, Tipalti Detect
  • Our machine learning-powered solutions provide real-time detection to help users validate accounts securely and prevent fraudulent transactions.23
    Trustpair, Trustpair

Source: the makers' own pages, read 29 October 2026; notes 10 to 23.

3Limits and control questions

For a listed company, any automated step that feeds the financial statements sits inside internal control over financial reporting, which management must evaluate and report on; the SEC's interpretive guidance for that report is Release 33-8810.4 COSO describes its Internal Control, Integrated Framework as originally issued in 1992 and refreshed in 2013.5

Thresholds are a judgment, not a setting. The SEC staff's view on materiality is that "exclusive reliance on this or any percentage or numerical threshold has no basis in the accounting literature or the law."6 A fraudulent payment can be small; a model tuned only to large anomalies will miss it.

Fraud models need feedback. Ask how confirmed fraud and false alarms are fed back, and whether the model learns from other customers' confirmed cases.

NIST's AI Risk Management Framework is a common reference for those questions. NIST says the framework is "intended for voluntary use and to improve the ability to incorporate trustworthiness considerations into the design, development, use, and evaluation of AI products, services, and systems."7 Its first version is published as NIST AI 100-1.8

Vendor claims about AI are claims, not findings. The SEC has brought settled charges against two investment advisers "for making false and misleading statements about their purported use of artificial intelligence (AI)."9 A finance team that relies on a vendor's description should ask for evidence of what the feature does with its own data.

4What filers say

A full-text search of SEC filings made between June 2025 and October 2026 found no filer describing AI in its own work on this task in words specific enough to quote. See how filings were searched.

Exhibit 11.2.Finance-software filers on their AI products

Listed software companies describing their own AI products to investors. These are the makers' statements, made in filings.

  • Audit resolution Workflow supported by AI utilizing our contract management platform as the primary data source.26
    CASS INFORMATION SYSTEMS INC (CASS), Form 8-K filed 23 April 2026
  • Our P2P solution also records receipt of goods and invoices and performs AI-based three-way matching.27
    XBP Global Holdings, Inc. (XBP, XBPEW), Form 10-K filed 31 March 2026

Source: SEC EDGAR full-text search; notes 26 to 27.

5Products

Exhibit 11.3.Products addressing anomaly and fraud detection in payments, by the data they name

12 product records whose makers say they address this task, alphabetical by maker. Open a row for our summary, the data it names and the maker's words on review; the marks show which questions the maker's pages answer.

  1. AI in BILLBILLData it names: not stated.Where a person reviews: quoted.Model or data handling: quoted.
    Our summary
    AI features across an AP and AR platform for smaller and mid-sized businesses, including anomaly and fraud checks on bills. The maker says the customer keeps final decisions.
    Data it names (our list)
    Not stated on the pages read.
    Where a person reviews (quoted)
    while keeping final decision-making power.
  2. Payments Fraud DefenseBottomlineData it names: quoted.Where a person reviews: not stated.Model or data handling: not stated.
    Our summary
    A cloud fraud prevention platform that screens payments in real time, sold to banks and integrated into their payment systems. The page names AI integration as part of the fraud strategy.
    Data it names (our list)
    Bank payment systems
    Where a person reviews (quoted)
    Not stated on the pages read.
  3. Coupa AI, Coupa Navi agentsCoupaData it names: not stated.Where a person reviews: not stated.Model or data handling: quoted.
    Our summary
    AI in Coupa's spend management platform for detecting fraud, policy violations and supplier risks, including AI-powered fraud detection and PO matching in accounts payable. Coupa Navi agents retrieve documents, speed approvals and answer questions.
    Data it names (our list)
    Not stated on the pages read.
    Where a person reviews (quoted)
    Not stated on the pages read.
  4. EftsureEftsureData it names: not stated.Where a person reviews: quoted.Model or data handling: not stated.
    Our summary
    Payment verification software that checks payee identity and bank account ownership before funds move, with machine-learning models that scan payment files. Fraud analysts review cases that automated checks cannot resolve.
    Data it names (our list)
    Not stated on the pages read.
    Where a person reviews (quoted)
    Trained fraud analysts investigate and make independent, out-of-band callbacks whenever automated checks can't resolve a case.
  5. Kyriba Liquidity Performance PlatformKyribaData it names: not stated.Where a person reviews: not stated.Model or data handling: quoted.
    Our summary
    AI in Kyriba's treasury platform for cash forecasting and scenario modelling, with real-time anomaly detection on payments. The maker says every AI insight is explainable and auditable.
    Data it names (our list)
    Not stated on the pages read.
    Where a person reviews (quoted)
    Not stated on the pages read.
  6. Medius AI, Medius CopilotMediusData it names: not stated.Where a person reviews: not stated.Model or data handling: not stated.
    Our summary
    AI across AP for invoice capture, coding, supplier questions and payment batches, with anomaly and fraud checks. The maker says payment batch recommendations carry a reason and source.
    Data it names (our list)
    Not stated on the pages read.
    Where a person reviews (quoted)
    Not stated on the pages read.
  7. MindBridge, Autonomous Financial OversightMindBridgeData it names: not stated.Where a person reviews: not stated.Model or data handling: not stated.
    Our summary
    Analytics that score the full population of financial transactions for unusual activity, used by finance and audit teams. The maker says it explains what drives each risk.
    Data it names (our list)
    Not stated on the pages read.
    Where a person reviews (quoted)
    Not stated on the pages read.
  8. Travel and expense monitoringOversightData it names: quoted.Where a person reviews: not stated.Model or data handling: quoted.
    Our summary
    Risk monitoring for travel and expense and card spend that prioritises high-risk items for auditors. The maker says its models are trained on anonymised transactions and auditor decisions.
    Data it names (our list)
    SAP Concur, Workday, Emburse, Mastercard, Bank card issuers
    Where a person reviews (quoted)
    Not stated on the pages read.
  9. Safebooks AI agentsSafebooks AIData it names: quoted.Where a person reviews: quoted.Model or data handling: not stated.
    Our summary
    AI agents that run finance processes such as purchase-to-pay, revenue and the close on a connected record of financial data. The maker says exceptions go to a named person, who decides.
    Data it names (our list)
    ERP, CRM, Billing, HRIS, Banks, Contracts
    Where a person reviews (quoted)
    The agent presents the context, source documents and recommended action; the human decides, and the agent records the decision.
  10. Billy, Stampli AIStampliData it names: quoted.Where a person reviews: quoted.Model or data handling: not stated.
    Our summary
    An AI assistant in procure-to-pay software that codes invoices, checks lines, routes approvals and flags changed vendor payment details. The maker says humans confirm what it suggests.
    Data it names (our list)
    Sage Intacct, Oracle NetSuite, SAP, Microsoft Dynamics, Acumatica, Sage, QuickBooks, Dealertrack
    Where a person reviews (quoted)
    Billy follows your approval rules and calls attention to exceptions that need judgment.
  11. Tipalti AI, AP Agent, Tipalti DetectTipaltiData it names: quoted.Where a person reviews: quoted.Model or data handling: quoted.
    Our summary
    AI in AP software from invoice capture to ERP sync, with field-level reasoning for recommendations, and a payee fraud screening feature. The maker says customer data does not train third-party models.
    Data it names (our list)
    ERP accounting data
    Where a person reviews (quoted)
    Understand recommendations with field-level reasoning before accepting them
  12. TrustpairTrustpairData it names: quoted.Where a person reviews: not stated.Model or data handling: not stated.
    Our summary
    Vendor fraud prevention software that validates suppliers' bank accounts and analyses payment files for suspicious behaviour before payment. Its home page says Basware has acquired it.
    Data it names (our list)
    SAP, Oracle
    Where a person reviews (quoted)
    Not stated on the pages read.

Source: the makers' own pages, read 29 October 2026; notes 10 to 25.

6Questions to put to a vendor

  1. Which signals does the product check before a payment is released, and which are rules and which are models?
  2. How does it detect a change of bank details, and what does it require before the change is accepted?
  3. Can a held payment be released by the same person who entered it?
  4. What share of alerts were false alarms in a pilot on our data, and how long did review take?
  5. Does it use confirmed fraud patterns from other customers, and how is that data protected?
  6. How does it handle documents or voices that may be generated, such as deepfake approvals?
  7. What does it log for each alert, decision and release?
  8. How does it fit with our bank's own payment screening, and who is responsible for which check?

Notes

  1. Federal Bureau of Investigation, Internet Crime Complaint Center, Internet Crime Report 2024. www.ic3.gov/AnnualReport/Reports/2024_IC3Report.pdf. Read 29 October 2026.
  2. Financial Crimes Enforcement Network, FinCEN Alert on Fraud Schemes Involving Deepfake Media Targeting Financial Institutions. www.fincen.gov/sites/default/files/shared/FinCEN-Alert-DeepFakes-Alert508FINAL.pdf. Read 29 October 2026.
  3. Internal Revenue Service, Taxpayer Identification Number (TIN) Matching. www.irs.gov/tax-professionals/taxpayer-identification-number-tin-matching. Read 29 October 2026.
  4. U.S. Securities and Exchange Commission, Commission Guidance Regarding Management's Report on Internal Control Over Financial Reporting, Release 33-8810. www.sec.gov/rules/interp/2007/33-8810.pdf. Read 29 October 2026.
  5. COSO, Internal Control, Integrated Framework. www.coso.org/guidance-on-ic. Read 29 October 2026.
  6. U.S. Securities and Exchange Commission, Staff Accounting Bulletin No. 99: Materiality. www.sec.gov/interps/account/sab99.htm. Read 29 October 2026.
  7. National Institute of Standards and Technology, AI Risk Management Framework. www.nist.gov/itl/ai-risk-management-framework. Read 29 October 2026.
  8. National Institute of Standards and Technology, Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1. nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf. Read 29 October 2026.
  9. U.S. Securities and Exchange Commission, SEC Charges Two Investment Advisers with Making False and Misleading Statements About Their Use of Artificial Intelligence. www.sec.gov/newsroom/press-releases/2024-36. Read 29 October 2026.
  10. BILL, Increase Efficiency & Accuracy with AI in BILL. www.bill.com/product/ai. Read 29 October 2026.
  11. Bottomline, Payments Fraud Defense | Bottomline. www.bottomline.com/us/fraud-compliance/payments-fraud-defense. Read 29 October 2026.
  12. Coupa, AI-Native Agentic Spend Management Platform. www.coupa.com/platform/. Read 2 November 2026.
  13. Coupa, Discover the Coupa AI Difference. www.coupa.com/platform/ai/. Read 2 November 2026.
  14. Eftsure, Eftsure. www.eftsure.com/. Read 2 November 2026.
  15. Kyriba, Kyriba Liquidity Performance Platform. www.kyriba.com/products/liquidity-performance/. Read 2 November 2026.
  16. Medius, Autonomous Accounts Payable | Medius. www.medius.com/ai-innovation/autonomous-accounts-payable/. Read 30 October 2026.
  17. MindBridge, MindBridge AI | The Leader in Autonomous Financial Oversight. www.mindbridge.ai/. Read 30 October 2026.
  18. Oversight, AI-powered Travel & Expense Monitoring | Oversight. oversight.com/ai-travel-expense-monitoring. Read 31 October 2026.
  19. Safebooks AI, AI Agents for Finance: How They Work | Safebooks AI. safebooks.ai/ai-agents-for-finance/. Read 2 November 2026.
  20. Stampli, AI for Procure-to-Pay Automation | Billy by Stampli. www.stampli.com/ai/. Read 1 November 2026.
  21. Tipalti, Tipalti AI | AI Agents for Finance Teams. tipalti.com/accounts-payable-software/finance-ai/. Read 1 November 2026.
  22. Tipalti, Payment Fraud Detection Software: Strengthen Your Defense. tipalti.com/accounts-payable-software/payment-fraud-detection/. Read 1 November 2026.
  23. Trustpair, Trustpair platform. trustpair.com/platform/. Read 2 November 2026.
  24. Stampli, AI Information - Stampli. www.stampli.com/ai-information. Read 1 November 2026.
  25. Trustpair, Trustpair. trustpair.com/. Read 2 November 2026.
  26. SEC EDGAR, CASS INFORMATION SYSTEMS INC, Form 8-K filed 23 April 2026. www.sec.gov/Archives/edgar/data/708781/000070878126000020/investorpresentation2026.htm. Read 29 October 2026.
  27. SEC EDGAR, XBP Global Holdings, Inc., Form 10-K filed 31 March 2026. www.sec.gov/Archives/edgar/data/1839530/000110465926037421/xbp-20251231x10k.htm. Read 29 October 2026.

Other use cases: Forecasting, Accounts payable, Expense audit, Collections and cash application, Reconciliation, Close automation, Variance analysis, Contract and lease data, Tax research, Treasury and FX, Audit preparation.