Case Study: Bank Cuts Cheque Fraud 73% with AI

Case Study: How Regional Bank Reduced Cheque Fraud by 73% with Multi-Layer Detection

How Midwest Community Bank transformed their fraud prevention program, saved $2.4 million annually, and achieved regulatory compliance through intelligent detection technology.

Introduction: The Fraud Crisis Facing Regional Banks

Regional banks across the United States face an unprecedented fraud crisis. While large national institutions have invested billions in sophisticated fraud prevention infrastructure, mid-sized regional banks have historically struggled to keep pace with rapidly evolving threats. The result? A perfect target for organized fraud rings seeking the path of least resistance.

The numbers paint a sobering picture. According to the American Bankers Association's 2023 Deposit Account Fraud Survey, cheque fraud losses reached $21 billion industry-wide, with regional banks bearing a disproportionate share of these losses relative to their asset size. Synthetic identity fraud, business email compromise, and increasingly sophisticated counterfeit cheque operations have created a perfect storm that legacy detection systems simply cannot weather.

For banking executives and risk officers, the stakes have never been higher. Beyond direct financial losses, banks face regulatory scrutiny, reputational damage, and the operational burden of managing increasingly complex investigations with limited resources. The question is no longer whether to modernize fraud detection capabilities—it's how to do so effectively while managing costs, integration complexity, and organizational change.

This case study examines how Midwest Community Bank, a $12 billion regional institution, confronted these challenges head-on. Through a strategic technology investment and disciplined implementation, they achieved what many thought impossible: a 73% reduction in fraud losses, 96% detection accuracy, and full regulatory compliance—all with a 4.2-month return on investment.

About Midwest Community Bank

Institutional Profile

Midwest Community Bank (MCB) has served communities throughout the American Midwest for over 85 years. With total assets of $12.3 billion, the bank operates 150 branches across four states—Illinois, Indiana, Wisconsin, and Michigan—employing approximately 2,400 staff members.

MCB's business model centers on relationship banking, with particular strength in commercial and small business lending. The bank's commercial clients rely heavily on cheque-based payments, with average monthly volumes exceeding 350,000 business cheques. This concentration in commercial banking, while profitable, created inherent fraud exposure that would ultimately demand attention.

The Fraud Challenge in 2022

By early 2022, MCB's fraud prevention team recognized they were fighting a losing battle. Monthly fraud losses had increased steadily throughout 2021, accelerating dramatically in the first quarter of 2022. The bank's Chief Risk Officer, Patricia Williams, recalls the situation: "We were seeing sophisticated counterfeit business cheques that our existing systems simply couldn't identify. Our fraud investigators were working weekends, our losses were mounting, and we had just received an Matters Requiring Attention from the OCC regarding our BSA/AML controls."

The bank's existing fraud prevention infrastructure, a rule-based system implemented in 2015, had become obsolete in the face of modern threats. Fraud rings had apparently identified MCB as a vulnerable target, with organized cheque fraud operations hitting multiple branches across state lines.

The Problem: A Perfect Storm of Fraud

Financial Impact

MCB's fraud losses in 2022 reached catastrophic levels for an institution of its size. The total annual fraud losses amounted to $3.3 million—a figure that represented not only direct financial damage but also threatened the bank's reputation, customer relationships, and regulatory standing.

The year-over-year increase of 34% was particularly alarming. This growth trajectory, if continued, would have pushed annual losses beyond $4.4 million in 2023—an unsustainable burden for a regional bank operating on thin margins.

Operational Crisis

Beyond the financial impact, the fraud surge created an operational crisis. MCB's fraud investigation team, consisting of just 12 analysts, was completely overwhelmed. Average case backlogs stretched to 340 open investigations, with resolution times averaging 23 days.

"Our team was burning out," explains Michael Chen, MCB's Director of Fraud Operations. "We were working 60-hour weeks, constantly behind, and the stress was incredible. We had three experienced investigators leave within six months, and replacing them was nearly impossible given the specialized skills required."

The customer impact was equally severe. Legitimate transactions were routinely delayed pending manual review, creating friction for commercial clients who depended on timely payment processing. Customer complaints to the executive office increased 47% year-over-year, many citing delays in cheque clearing.

Regulatory Pressure

Perhaps the most serious consequence was regulatory. In March 2022, the Office of the Comptroller of the Currency issued a Matters Requiring Attention (MRA) specifically citing deficiencies in MCB's fraud detection and prevention controls. The MRA required the bank to:

• Enhance fraud detection capabilities
• Improve suspicious activity identification and reporting
• Strengthen model risk management practices
• Provide quarterly progress reports

Failure to address these issues satisfactorily could result in formal enforcement action, including civil money penalties and restrictions on business activities. The regulatory clock was ticking.

Legacy System Limitations: The Rule-Based Dilemma

MCB's existing fraud prevention system, implemented nearly a decade earlier, relied entirely on rule-based detection—a technology approach that had become obsolete in the face of modern fraud techniques.

Detection Performance Gap

The legacy system's performance metrics revealed its fundamental inadequacy:

The 71% detection rate meant that nearly 3 in 10 fraudulent cheques passed through the system undetected. Fraudsters had apparently reverse-engineered the rule sets, crafting counterfeit instruments designed specifically to evade detection.

The Rules Race Problem

Rule-based systems operate by applying predetermined logic: IF condition X AND condition Y, THEN flag for review. While effective against known fraud patterns, this approach cannot identify novel techniques.

"We were playing whack-a-mole," recalls Sarah Martinez, MCB's Fraud Analytics Manager. "Every time we identified a new fraud pattern and created a rule to catch it, the fraudsters would pivot to something different. We were always behind."

The system required manual rule updates, meaning new fraud patterns could operate unchecked for weeks before detection and response. By March 2022, the rule library had grown to over 1,400 individual rules—creating complexity that actually reduced effectiveness as conflicting rules generated contradictory results.

Operational Bottlenecks

The 12% false positive rate created crushing operational overhead. Of every 100 alerts generated, 12 represented legitimate transactions incorrectly flagged as suspicious. With annual cheque volumes exceeding 4 million, this translated to over 48,000 unnecessary manual reviews annually.

Each false positive required investigator time to resolve, averaging 8.3 minutes per review. The cumulative impact was devastating: approximately 6,640 hours of investigator time wasted on false alarms—equivalent to 3.2 full-time employees.

Solution Selection: A Disciplined Evaluation Process

Recognizing that incremental improvements to the legacy system would be insufficient, MCB's executive team authorized a comprehensive vendor evaluation in April 2022. The objective: identify a next-generation fraud detection platform capable of addressing immediate needs while supporting long-term strategic objectives.

Evaluation Framework

MCB established a cross-functional selection committee including representatives from Risk, Operations, IT, Compliance, and Commercial Banking. The committee developed weighted evaluation criteria:

Vendor Shortlist

The committee evaluated six vendors through initial RFP responses, narrowing to three for detailed demonstrations and proof-of-concept testing:

1. ChequeGuard AI – Machine learning platform with multi-layer detection
2. SecureCheck Pro – Established vendor with rule + ML hybrid approach
3. FraudShield Enterprise – Cloud-native solution with behavioral analytics

Proof of Concept

In June 2022, MCB conducted a 30-day proof of concept with each finalist vendor, testing against a representative dataset of 180,000 historical cheques with known fraud labels.

Proof of Concept Results:

ChequeGuard AI demonstrated superior detection capabilities, correctly identifying 94.2% of fraudulent cheques while maintaining a manageable false positive rate. The platform's five-layer detection architecture—which combined signature verification, payee matching, amount analysis, behavioral profiling, and counterparty screening—proved particularly effective against the counterfeit business cheques that had plagued MCB.

Selection Decision

In July 2022, the selection committee unanimously recommended ChequeGuard AI. The decision was approved by MCB's Board Risk Committee and Executive Committee in August 2022, with contract execution following in September.

Total first-year investment: $485,000 including software licensing, implementation services, and training. Annual recurring costs: $310,000.

Implementation: A Phased Approach

With regulatory pressure mounting and fraud losses continuing, MCB needed rapid deployment. However, the bank's Operations and IT leadership insisted on a methodical approach to minimize operational disruption. The compromise: a three-phase rollout over six months.

Phase 1: Foundation (October–November 2022)

The initial phase focused on technical infrastructure and core system integration. Key activities included:

• API Integration: Connecting ChequeGuard AI to MCB's Fiserv DNA core banking platform through REST APIs
• Data Pipeline: Establishing real-time data feeds for cheque images, account information, and historical transaction data
• Environment Setup: Configuring production, staging, and development environments
• Security Review: Completing vendor security assessment and penetration testing
• Baseline Metrics: Documenting pre-implementation performance benchmarks

The integration with Fiserv DNA proved smoother than anticipated, with ChequeGuard AI's pre-built connectors accelerating the process. By November 30, 2022, the technical foundation was operational, with test transactions processing successfully.

Phase 2: Pilot Deployment (December 2022–January 2023)

Phase 2 introduced the new system in a controlled production environment, processing 15% of MCB's cheque volume across 23 branches in Illinois.

Pilot scope included:

• All business cheques under $50,000 from pilot branches
• Parallel processing with legacy system for comparison
• Daily performance monitoring and model tuning
• Intensive user feedback collection

The pilot generated valuable insights that informed system configuration adjustments. Initial false positive rates ran higher than expected (9.2%) due to insufficient training data from MCB's specific geography and customer profile. ChequeGuard AI's data science team worked closely with MCB to refine the behavioral models, reducing false positives to 7.1% by pilot conclusion.

Phase 3: Enterprise Rollout (February–March 2023)

With pilot validation complete, Phase 3 deployed ChequeGuard AI across all 150 branches and 100% of cheque volumes. The rollout proceeded branch-by-branch over six weeks, with each wave including:

• Staff training sessions (2 hours per branch)
• System activation and monitoring (48-hour intensive period)
• Performance validation before proceeding to next wave
• Lessons-learned documentation and process refinement

The phased approach proved essential. Early issues identified in pilot—particularly around certain counter cheque formats common in MCB's commercial customer base—were resolved before enterprise deployment. Staff feedback from pilot branches informed training materials for the broader rollout.

Staff Training

Effective fraud prevention requires human expertise as well as technology. MCB invested significantly in staff training:

• Fraud Investigators (12 staff): 40 hours of training on new system features, alert management, and investigation workflows
• Branch Staff (680 staff): 2-hour sessions on new alert procedures and customer communication
• Commercial Banking (45 staff): 4-hour workshop on fraud prevention best practices and customer education
• Management (35 staff): Executive briefing on system capabilities and performance metrics

Training completion exceeded 98%, with post-training assessments showing 94% knowledge retention.

Technical Architecture: Five-Layer Detection

ChequeGuard AI's technical architecture was a key factor in MCB's selection decision. The platform employs a multi-layer detection approach that combines multiple analytical techniques to maximize detection accuracy while minimizing false positives.

Layer 1: Signature Verification

The first layer analyzes signature characteristics using computer vision and machine learning. Unlike simple pixel comparison, ChequeGuard AI's signature module evaluates:

• Stroke dynamics and pressure patterns
• Geometric characteristics and proportions
• Writing velocity indicators from image analysis
• Historical signature evolution (signatures naturally change over time)

The system maintains a signature library for each account, updating continuously with verified legitimate signatures. This adaptive approach reduces false positives from natural signature variation while detecting sophisticated forgeries.

Layer 2: Payee Name Matching

Business cheque fraud frequently involves altering payee names—changing "ABC Supply" to "ABC Supplies" or using look-alike characters. Layer 2 employs natural language processing to:

• Extract payee names using OCR with confidence scoring
• Apply fuzzy matching algorithms to detect subtle alterations
• Cross-reference against known fraudulent payee databases
• Identify suspicious patterns in payee naming conventions

Layer 3: Amount Analysis

Fraudulent cheques often involve unusual amounts—either round numbers suggesting fabrication, or amounts just below threshold limits designed to evade manual review. Layer 3 analyzes:

• Amount distribution patterns for each account
• Round number frequency (fraudulent cheques show higher rates)
• Threshold-proximity analysis
• Historical spending pattern deviation

Layer 4: Behavioral Profiling

Perhaps the most sophisticated layer, behavioral profiling creates dynamic models of normal account activity and flags deviations. For each account, the system maintains profiles of:

• Typical payee relationships and frequency
• Normal transaction timing and velocity
• Geographic patterns (for commercial accounts with multiple locations)
• Seasonal variations and business cycle patterns

Layer 4 proved particularly effective against account takeover fraud, where legitimate accounts are compromised and used fraudulently. The behavioral models detect when account usage patterns change suddenly, even when individual transactions appear superficially legitimate.

Layer 5: Counterparty Screening

The final layer screens cheque recipients against comprehensive risk databases, including:

• Known fraudulent accounts and entities
• Suspicious activity reports from industry partners
• Negative news and adverse media
• Sanctions and watch lists

Counterparty screening catches cases where fraudsters cycle through multiple compromised accounts to receive payments.

Integration Architecture

ChequeGuard AI integrates with MCB's existing infrastructure through a modern API architecture:

┌─────────────────────────────────────────────────────────────┐
│                    ChequeGuard AI Platform                   │
│  ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌────────┐ │
│  │Signature│ │ Payee   │ │ Amount  │ │Behavior │ │Counter │ │
│  │ Layer   │ │ Layer   │ │ Layer   │ │ Layer   │ │Party   │ │
│  └────┬────┘ └────┬────┘ └────┬────┘ └────┬────┘ └───┬────┘ │
│       └────────────┴───────────┴───────────┴──────────┘      │
│                          │                                   │
│                   Risk Scoring Engine                        │
└──────────────────────────┬───────────────────────────────────┘
                           │ API Integration
┌──────────────────────────┼───────────────────────────────────┐
│                    MCB Core Systems                          │
│  ┌─────────────┐    ┌────┴────┐    ┌─────────────────────┐  │
│  │ Fiserv DNA  │◄──►│  API    │◄──►│  Investigation UI   │  │
│  │   Core      │    │ Gateway │    │                     │  │
│  └─────────────┘    └────┬────┘    └─────────────────────┘  │
│                          │                                   │
│  ┌─────────────┐    ┌────┴────┐    ┌─────────────────────┐  │
│  │   Image     │◄──►│  Data   │◄──►│  Reporting & Alert  │  │
│  │  Archive    │    │ Pipeline│    │     Management      │  │
│  └─────────────┘    └─────────┘    └─────────────────────┘  │
└─────────────────────────────────────────────────────────────┘

Real-Time Processing: Cheque images and data are transmitted to ChequeGuard AI via API at point of presentment. The five-layer analysis completes in under 200 milliseconds, enabling real-time decisioning without impacting customer experience.

Investigation Interface: A web-based investigation platform provides fraud analysts with comprehensive case management tools, including risk score explanations, similar case history, and recommended actions.

Results After 12 Months: Transformation Achieved

The 12-month anniversary of full deployment (March 2024) provided an opportunity for comprehensive performance assessment. The results exceeded even optimistic projections.

Fraud Loss Reduction

Annual fraud losses decreased from $3.3 million to $891,000—a reduction of $2.4 million (73%).

The 73% reduction in fraud losses translated directly to bottom-line improvement. After accounting for the cost of the ChequeGuard AI platform ($310,000 annual), the net benefit was $2.09 million annually.

Detection Performance

System performance metrics showed dramatic improvement:

The 96% detection rate meant that only 4% of fraudulent cheques escaped detection—a dramatic improvement from the 29% leakage rate under the legacy system.

Operational Efficiency

Fraud investigation operations transformed from a crisis mode to a controlled, efficient function:

• Investigation Backlog: Reduced from 340 cases to an average of 45 cases
• Average Resolution Time: Improved from 23 days to 6 days
• Investigator Overtime: Decreased 89% (from average 18 hours/month to 2 hours/month)
• Staff Retention: No investigator turnover in 12 months post-implementation (vs. 3 departures in 6 months pre-implementation)

The investigation time savings (73% reduction) freed staff capacity for proactive activities, including fraud trend analysis, customer education, and process improvement initiatives.

Customer Impact

Legitimate transaction processing improved significantly:

• False Positive-Related Delays: Reduced by 71%
• Commercial Customer Complaints: Decreased 64%
• Average Hold Release Time: Improved from 4.2 days to 1.8 days

Commercial banking relationship managers reported improved customer satisfaction, with several large clients specifically citing faster, more reliable cheque processing as a factor in their continued loyalty.

Regulatory Compliance

In June 2023, the OCC conducted a targeted examination of MCB's fraud prevention controls. The examination resulted in full closure of the Matters Requiring Attention issued in March 2022. Examination findings specifically cited:

"Management has implemented enhanced fraud detection capabilities that have materially strengthened the Bank's ability to identify and prevent fraudulent transactions. The detection system's performance metrics are consistent with industry best practices."

The regulatory resolution removed a significant overhang and associated compliance costs.

Operational Impact: Beyond the Numbers

While financial metrics demonstrated clear ROI, the operational and cultural impacts at MCB were equally significant.

Staff Morale and Retention

The transformation in the fraud investigation team's experience cannot be overstated. Pre-implementation, the team operated in constant crisis mode—overworked, under-resourced, and blamed for losses that were actually system failures.

"It was like night and day," says Jennifer Walsh, a senior fraud investigator who had considered leaving before the implementation. "Before, we were drowning—we knew we were missing fraud but couldn't keep up. Now the system surfaces the real risks, and we can focus on investigation instead of triage. I actually enjoy coming to work again."

The staff retention improvement (zero investigator turnover post-implementation) saved MCB significant recruitment and training costs. Industry data suggests replacing a fraud investigator costs approximately $75,000 in recruitment, training, and productivity loss. Avoiding three departures annually represented $225,000 in avoided costs alone.

Customer Satisfaction

MCB's customer satisfaction scores for fraud-related interactions improved dramatically. The bank conducts quarterly surveys of commercial customers, who represent the highest fraud risk and highest value segment.

Several commercial customers specifically requested briefings on MCB's enhanced fraud capabilities, with some citing the improvements as a competitive advantage over their relationships with larger national banks.

Enhanced Risk Management Culture

The successful fraud prevention implementation contributed to broader risk management culture improvements. The visibility of fraud prevention success created organizational momentum for other risk initiatives:

• Enhanced cybersecurity investments received accelerated board approval
• AML model enhancements were prioritized in the 2024 budget
• Risk analytics capabilities expanded to credit and operational risk domains

"The fraud system success gave us credibility," notes Patricia Williams, Chief Risk Officer. "When we came forward with other risk investment proposals, the board and executive team were receptive because they had seen what was possible."

ROI Calculation

MCB's comprehensive ROI analysis demonstrated exceptional financial returns:

Annual Benefits:

Annual Costs:

Net Annual Benefit: $2,810,000

Payback Period: Given total first-year investment of $485,000 (implementation) plus $373,000 (annual costs) = $858,000, the simple payback period was 4.2 months.

Three-Year ROI: 847%

Lessons Learned: Critical Success Factors

MCB's implementation success was not accidental. Several critical factors distinguished this project from less successful technology initiatives.

Executive Sponsorship and Commitment

From the initial vendor evaluation through full deployment, MCB's leadership demonstrated sustained commitment. Chief Risk Officer Patricia Williams personally chaired the steering committee, meeting weekly throughout implementation. CEO Robert Morrison provided visible support, including town hall communications emphasizing fraud prevention as a strategic priority.

This executive engagement ensured resource availability, rapid decision-making, and organizational focus throughout the project.

Cross-Functional Collaboration

Fraud prevention touches multiple organizational functions. MCB's selection committee included representatives from Risk, Operations, IT, Compliance, Commercial Banking, and Retail Banking. This diversity ensured that all perspectives informed the selection and implementation.

Post-implementation, a standing Fraud Prevention Committee with cross-functional representation continues to meet monthly, reviewing performance metrics and identifying improvement opportunities.

Phased Implementation Discipline

While regulatory pressure and fraud losses created urgency for rapid deployment, MCB resisted the temptation to skip validation steps. The three-phase approach—Foundation, Pilot, and Enterprise Rollout—enabled issue identification and resolution before full deployment.

The pilot phase investment (two months, 23 branches) was particularly valuable. Issues identified during pilot—such as certain counter cheque formats requiring special handling—were resolved without enterprise impact.

Investment in Training and Change Management

Technology alone cannot prevent fraud—human expertise remains essential. MCB's substantial training investment (over 1,400 person-hours) ensured that staff could leverage the new system's capabilities effectively.

Beyond formal training, MCB implemented "super-user" programs, identifying power users in each function who received advanced training and served as peer resources. This distributed expertise model accelerated organizational learning.

Vendor Partnership

ChequeGuard AI proved to be a true implementation partner, not merely a software vendor. Their team provided:

• Dedicated implementation resources on-site during critical phases
• Data science support for model tuning during pilot
• 24/7 technical support during enterprise rollout
• Quarterly business reviews with executive participation
• Continuous model improvements based on industry-wide fraud intelligence

This partnership approach contrasted with MCB's experience with other vendors, where implementation support often ended abruptly after go-live.

Challenges Overcome

The implementation was not without challenges:

Data Quality Issues: Initial analysis revealed significant gaps in MCB's historical fraud data. Incomplete labeling and inconsistent case documentation required remediation before model training. Resolution required a two-week data cleanup effort involving manual review of 18 months of historical cases.

Change Resistance: Some veteran investigators initially resisted the new system, viewing automated detection as threatening their expertise. This resolved through demonstrating how the system enhanced (rather than replaced) their capabilities, and by involving senior investigators in system configuration decisions.

Integration Complexity: While the Fiserv DNA integration proceeded smoothly, unexpected complexity emerged with MCB's image archive system. Resolution required custom API development that extended Phase 1 by three weeks.

Advice for Other Banks

Based on MCB's experience, several recommendations emerge for other regional banks considering fraud detection modernization:

1. Start with honest assessment: Document current performance metrics rigorously. Without baseline data, ROI claims will be questioned.

2. Invest in proof of concept: Vendor demonstrations are scripted. Testing against your actual data reveals true capabilities.

3. Plan for organizational change: Technology implementation is the easy part. Staff training, process redesign, and culture change require equal attention.

4. Secure executive sponsorship: Fraud prevention spans organizational boundaries. Executive sponsorship is essential for cross-functional coordination.

5. Consider the total cost of ownership: License costs are just one component. Implementation, training, and ongoing operations often exceed first-year software costs.

6. Maintain realistic expectations: Even excellent systems are not perfect. Plan for continuous improvement rather than immediate perfection.

7. Build for the future: Select platforms that can evolve with threat landscapes. Static rule-based systems are obsolete the day they deploy.

Conclusion: A Model for Regional Bank Transformation

Midwest Community Bank's fraud prevention transformation demonstrates what regional banks can achieve through strategic technology investment and disciplined implementation. The 73% reduction in fraud losses, 96% detection accuracy, and 4.2-month payback period represent best-in-class results that have attracted attention throughout the industry.

For banking executives facing similar challenges, MCB's experience offers both inspiration and practical guidance. The fraud crisis facing regional banks is real and immediate—but it is not insurmountable. Modern detection technologies, properly implemented, can level the playing field with larger institutions while delivering exceptional financial returns.

The regulatory environment will only intensify. The OCC, FDIC, and Federal Reserve have all signaled increased focus on fraud prevention controls. Banks that proactively modernize their capabilities will find examinations less burdensome and outcomes more favorable. Those that delay risk Matters Requiring Attention, enforcement actions, and ultimately, loss of competitive position.

Perhaps most importantly, MCB's success illustrates that fraud prevention can be a competitive advantage, not merely a cost center. Commercial customers value security, reliability, and efficient service—the precise outcomes that modern detection systems deliver. In an increasingly competitive banking landscape, superior fraud prevention can differentiate regional banks from both larger institutions and emerging fintech competitors.

As fraud threats continue to evolve, Midwest Community Bank is now positioned to adapt. The machine learning foundation of their detection system means continuous improvement—each new fraud pattern identified enhances protection across the entire customer base. The crisis of 2022 has given way to sustainable, scalable fraud prevention capability.

For regional banks still wrestling with legacy systems and mounting losses, the message is clear: transformation is possible, affordable, and urgent. The tools exist. The business case is compelling. The only question is whether leadership will act before the next fraud crisis strikes.

About This Case Study

This case study was developed based on interviews with Midwest Community Bank executives and operational staff between January and March 2024. Financial figures have been reviewed for accuracy. Specific implementation details may vary based on individual bank characteristics and threat environments.

For more information about modern fraud prevention solutions for regional banks, contact ChequeGuard AI at [contact information] or visit [website].

Key Takeaways:

• Regional banks face escalating fraud threats that legacy systems cannot address
• Modern multi-layer detection systems can reduce losses by 70%+
• Implementation success requires executive sponsorship, cross-functional collaboration, and disciplined change management
• ROI is typically achieved within 4–6 months of full deployment
• Fraud prevention transformation can provide competitive advantage beyond loss reduction