Skip to content

Anywhere On-Us Fraud

"Inclearing" and On-Us Check Fraud Detection and Prevention:                                                                                    Detecting Counterfeits, Forgeries, and Alterations at the Paying Bank 

Solution Overview: Anywhere On-Us Fraud

Anywhere On-Us Fraud is a specialized module within OrboGraph's comprehensive OrboAnywhere Suite , developed to detect and prevent On-Us check fraud at the paying bank. This module, when properly deployed and tuned, has proven to deliver 95%+ detection rates for targeted use cases.
On-Us checks are primarily received during the financial institution's inclearing process. Clearing items are vulnerable to alterations, counterfeits, and forgery because the original items were deposited at another bank or financial institution (BoFD) or cashed at a check casher. Liability for counterfeited checks and forged checks rests primarily on the paying bank, while alterations are the responsibility of the BoFD (although detectable at the Paying Bank). Even On-Us checks deposited or cashed at a paying bank are exposed to check fraud risk because many FI's do not have real-time check fraud detection at the teller or self service devices.
Anywhere On-Us Fraud colored-01

Anywhere On-Us Fraud Brochure

Complete the form below to download the Anywhere On-Us Fraud Brochure.

Anywhere Fraud - Onus

Anywhere On-Us Fraud Brochure
To effectively prevent check fraud attempts involving counterfeits, forgeries, and alterations, Anywhere On-Us Fraud employs a sophisticated combination of five (5) fraud detection layers of technology. These layers combine to deliver a best-in-breed check fraud detection solution. Additionally, the system generates hundreds of data attributes, which can be utilized in downstream business intelligence and data analytics systems.
The combination of advanced technology and comprehensive data analysis ensures that financial institutions are equipped with the necessary insights to safeguard their operations against evolving threats in check fraud.

Multi-layer Check Fraud Technology

A multi-layered check fraud technology approach is a strategic framework that leverages complementary technologies to create a more robust system. Anywhere Fraud emphasizes the integration of distinct components, each serving a specialized function while working cohesively within a larger ecosystem. By utilizing multiple layers, organizations minimize false positive rates (targeting under 0.5%) with low manual review overhead, while attaining 95%+ detection rates. 

Onus fraud

Layer 1: Infrastructure, Profiles, and Thresholds

Developing a robust multi-layered approach starts with a solid foundational infrastructure. By establishing a well-designed infrastructure, institutions can guarantee that data flows efficiently and risk scores are accurately calculated, paving the way for the most reliable detection system. 
Components of this layer include: multiple databases, APIs, Profile Editor user interface for threshold management, tuning, automated management, self learning profiles, and data import(s) tools. One of the pivotal components of on-us check fraud detection is an efficient and comprehensive account profiling system. System profiles start broad, then work their way into granular parameters. For example, a global profile may cover all retail checking accounts, while a small business profile, may target only small businesses, enabling finely tuned parameters for this customer type.
During the deployment and training phase for the Anywhere On-Us Fraud system, historical and transactional data—including images of previously cleared checks—are utilized to train account profiles. Once these profiles are established, financial institutions can set precise thresholds, employing both transactional and image forensic analyzers to pinpoint potential fraud very effectively. This careful calibration of thresholds significantly mitigates the number of false positives, allowing institutions to safeguard their operations while minimizing unnecessary disruptions to legitimate bank account activities. The self-learning profile logic is very intelligent, and self managing in profiling history, size, and storage.
Female Programmer Working in Monitoring Control Room, Surrounded by Big Screens Displaying Lines of Programming Language Code. Portrait of Woman Creating a Software. Abstract Futuristic Coding Concept

Layer 2: Outlier-Focused Transaction Analysis (Highly Targeted Data Analytics)

Transactional Analysis plays a crucial role in detecting check fraud by analyzing various meta data fields provided from checks, such as serial numbers and amounts. These statistical analyzers are designed and optimized specifically to identify anomalies that could indicate fraudulent activity, while learning the behaviors of the account.

Among the tools utilized are velocity analysis, which tracks the frequency of checks processed within a specific time frame, and out-of-range checks, including Serial Out of Range (SOOR) and Amount Out of Range (AOOR) analysis. These methods are pivotal for fraud analysis to flag unusual patterns that deviate from standard transaction behaviors, thus enabling institutions to spot potential fraudulent checks before they are honored.

Adding an additional layer of sophistication, these analyzers when combined with account-level profiling, enhance fraud detection capabilities. Another component of transactional analysis involves account status. If the account has a certain status, i.e. dormant, high risk, new, etc. the system can be setup to identify risks around each account type.

Analyst working with Business, Concept digital transformation technology strategy, Analytics and Data Management System on computer, make a report with KPI and metrics connected to database.

Lastly, the system considers businesses and consumers who have multiple check writers. A simple example would be a couple with both spouses writing checks. However, many large businesses issue checks from multiple offices, hence generating multiples of check serial numbers concurrently. Anywhere On-Us Fraud will dealing with these variations. For a full list of image and transaction analyzers, click the link below.

Transactional Analysis/Account Monitoring:

  • Serial Out of Range
  • Multiple Account Holder Profile (Serial)
  • Amount Out of Range
  • Duplicate Detection
  • Amount Thresholding
  • New Account Identification
  • Closed Account Identification
  • Inactive Account Identification
  • Watch List
  • Dormant Account
  • Real-Time Interface
Full List of Transactional Analyzers →

Image Forensics:

  • ASV-AI: Automated Signature Verification
  • CSV-AI: Check Stock Validation
  • WV-AI: Writer Verification
  • CAR/LAR Discrepancy: Alteration Detection
  • Check Style Analysis: Alteration Detection
  • Signature Presence Detection
  • Account Profiling (Self Learning)
  • Two Signatures Required
  • Preauthorized Draft Detection (PAD)
  • Fraudulent (“Known Bad”) Image Matching
  • ASV on Deposit Tickets
  • New Account Identification
  • Closed Account Identification
  • Inactive Account Identification
  • Watch List
  • Dormant Account
  • Real-Time Interface
Full List of Image Forensic Analyzers →

Layer 3: Image Forensics

Image forensics refers to the scientific analysis and examination of digital images of checks to identify signs of tampering, forgery, or manipulation. It involves using specialized techniques and tools to assess the authenticity and integrity of check images, ensuring they have not been altered to facilitate fraudulent activities.

Image forensics represents a significant evolution in image analysis technology, ushering in a new era where traditional methods like optical character recognition (OCR) and handwriting analysis algorithms are superseded by sophisticated artificial intelligence and deep learning models. This transition marks a fundamental shift in how image data is processed and analyzed, enabling greater accuracy and efficiency in detecting fraudulent activities. The current landscape requires more than basic analysis developed twenty years ago; it demands an adaptable approach, capable of learning from vast datasets to identify subtle changes and anomalies that may indicate fraud. 

Created by the OrboGraph OrbNet AI Innovation Lab, OrbNet Forensic AI goes beyond the constraints of legacy image analysis systems by employing deep learning. The recent release of OrboAnywhere Turbo delivers an important leap towards transparency in risk assessment. By introducing Explainable AI (XAI) , the new system provides explanatory result indicators which empower fraud analysts with clearer visibility into the risk scores generated. The support data attributes offer downstream data scientists additional optimization support. This combination of cutting-edge technology and enhanced transparency represents a transformative step forward in the realm of image forensics.

Suspect Item (Slide Right)

OrbNet Forensic AI Analysis On-Us Check

Companion Item

Digital looking check new fields-01
When dealing with check fraud detection, there are three most common technologies employed to identify the fraud type. They include:

Check Stock Validation (CSV-AI)

Check Stock Validation plays a crucial role in the detection of counterfeit checks by meticulously analyzing the attributes, layout, and relative coordinates of designated preprinted fields, which serve as “anchor points” on the check. This analytical process becomes increasingly vital in a landscape where fraudsters are leveraging advanced digital tools and generative AI to create sophisticated counterfeits, often referred to as “check cooking.” OrbNet Forensic AI steps in to combat this threat by identifying even the slightest inconsistencies between newly presented checks and previously cleared items. By evaluating each relevant field, the system assigns a risk score, offering a comprehensive overview of potential check fraud risks associated with the check.
The OrboAnywhere Turbo release includes advanced border and security symbol detection. These new features empower the system to conduct a more thorough examination of security elements that are critical to verify and compare a check's authenticity to previously cleared checks. Many fraudsters are unable to fully reproduce security symbols, enabling financial institutions to now build stronger defenses against these kinds of check fraud attacks. 

Automated Signature Verification (ASV-AI)

Automated Signature Verification is revolutionizing the way financial institutions authenticate signatures on checks, leveraging advanced AI technologies to ensure accuracy and security. Unlike traditional optical character recognition (OCR) systems , which rely on a rudimentary method of overlaying images to verify signatures, OrbNet Forensic AI employs a sophisticated forensic document examination approach. By activating 512 feature vectors specifically designed to analyze various attributes of signatures, this cutting-edge system compares the signer of a check against an extensive database of signatures captured from cleared checks on the same bank account . This comprehensive method significantly enhances the reliability of signature verification , reducing the risk of fraud and ensuring that only legitimate transactions are processed.

For additional insights into this innovative technology, interested readers can refer to the Forensic Document Examination interview, which offers a closer look at the implications and benefits of such advanced systems in safeguarding financial transactions.

Writer Verification (WV-AI)

As the prevalence of stolen checks continues to rise, fraudsters are employing increasingly sophisticated techniques to manipulate and alter these financial instruments. The process of “washing” checks—where original information is erased and replaced—has become alarmingly common, leaving many financial institutions struggling to safeguard their clients against such deceptive practices. In light of this escalating threat, Writer Verification has emerged as an innovative technology aimed at detecting altered checks through meticulous forensic analysis.

Writer Verification employs a comprehensive methodology to scrutinize various attributes of checks, including the Check Amount Recognition (CAR), Legal Amount Recognition (LAR), date, payee information, and comparisons to previously cleared items. By analyzing both printed text and handwritten elements, experts can identify subtle discrepancies that may indicate tampering. For instance, changes in the orientation, spacing, or slant of handwriting can reveal signs of alteration, while variations in printed fonts may also signify fraudulent intent. As financial institutions adopt these advanced verification techniques, they are better equipped to combat the growing threat of check fraud.

Alteration Detection

Alteration detection is a critical process in financial and data analysis, particularly when it comes to ensuring the integrity of financial reports and transactions. Two essential methods for identifying alterations are Check Style Analysis and the examination of courtesy and legal amount fields to identify discrepancies. These techniques are instrumental in uncovering discrepancies that may indicate manipulation or fraudulent activities. By scrutinizing these amount fields, analysts can effectively capture variations that may otherwise go unnoticed.

Layer 4: Rules Engine, Rules Creation, and Continuous Improvement

A fraud rules engine serves as a critical component in the arsenal of financial institutions striving to combat fraudulent activities. This specialized decision-making software utilizes a series of logic-based rules designed to identify potentially suspicious transactions or behaviors according to predefined parameters. By integrating advanced data analytics, financial institutions can analyze transaction patterns, enabling them to detect anomalies and establish new rules to address emerging fraud trends. This proactive approach is essential, as fraud schemes continuously evolve, necessitating a dynamic response from FIs to safeguard their operations and customers.

To enhance their fraud detection capabilities, Anywhere On-Us Fraud offers an embedded rules engine capable of managing a broader spectrum of fraud use cases. This sophisticated tool provides financial institutions with the flexibility to adjust the scoring of analyzers through the OA Rules Generator, a machine learning application developed by OrboGraph Client Services. By evaluating previously processed items, this application assists in formulating optimized business rules aimed at refining fraud detection mechanisms and minimizing false positives.

Capabilities involving a rules engine and rules creation help feed an OrboGraph process we call Continuous Improvement Process or CIP. By monitoring performance on an ongoing basis, we can fine tune rules and thresholds to continuously improve field production over time.

Document management concept, Businesswomen check electronic documents on digital documents on virtual screen Document Management System and process automation to efficiently document paperless operate

Layer 5: Fraud Review/Queues Companions & History

Technology plays a crucial role in identifying suspicious activities, but still requires highly skilled and well trained fraud experts to review "low score" items. Fraud analysts bring a depth of insight and experience. Equipped with the right tools and access to comprehensive data, these professionals are instrumental in deciphering complex fraud patterns and making informed decisions on potential threats.

To be truly effective in their roles, fraud analysts require sophisticated resources, such as a robust check fraud review platform, dedicated queues to assess alongside previously cleared transactional history, and companion images. Recognizing this need, Anywhere On-Us Fraud offers an integrated option to enhance the existing check fraud review platforms. A primary benefit is an API call for companion images. Furthermore, the incorporation explainable AI (XAI) offers analysts deeper insights into the rationale behind each fraud alert, enriching their investigative process. This combination of technology and expert analysis not only streamlines operations but also strengthens the overall defense against check fraud, ultimately safeguarding customer trust and financial assets.
Business,Professional,Looking,At,A,Fraud,Review,Platform,On,A

Solution Deployment and Benefits

Anywhere On-Us Fraud is leveraged by financial institutions of all sizes — from the top 10 banks to small community banks. The system is designed to be deployed across a variety of environments and workflows, catering to the specific needs and resources of the financial institution. Whether implemented on-premise for enhanced security control, in a service bureau, or in a public cloud for optimized performance and scalability, Anywhere On-Us Fraud adapts seamlessly to the technological landscape of financial institutions. Anywhere On-us Fraud is available for direct deployment, through business partners, or in combination with leading fraud review platforms (see business partners ).
Financial institutions of all sizes see the following benefits:
  • Increased detection of counterfeits, forgeries, and alterations
  • Reduction in false positives
  • Self-learning profiles provides continuous improvement
  • Flexible system deployment with existing fraud review platforms
  • Continous enhancements through product releases

Case Study
Top 15 Bank

Case Study

Profile: Top 10 Us Bank by Asset Size

Volume: 300,000+ items per day

Fraud $'s Caught (2024): $80,000,000+

Case Study
Regional Bank

Case Study

Profile: Regional Bank with over $35B in assets

Volume: 90,000+ items per day

Fraud $'s Caught (2024): $11,000,000+

Case Study
Community Bank

Case Study

Profile: Regional/Community

Volume: 40,000+ items per day

Fraud $'s Caught (2024): $6,000,000+

Detection Rate: 100%