How to tackle the challenges of AML monitoring

Banks, credit unions, and fintechs are expected to detect and report financial crime without slowing down good customers and members. That is hard. Transaction monitoring rules written years ago flag legitimate activity, analysts spend their days clearing alerts, and every delayed onboarding or held transaction risks a customer or member who never comes back.

By law, financial institutions must establish and maintain an Anti-Money Laundering (AML) program to detect and prevent money laundering, terrorist financing, and other illicit activity. The Bank Secrecy Act (BSA) of 1970 set reporting and recordkeeping requirements, including reporting cash transactions over $10,000. The USA PATRIOT Act of 2001 added customer due diligence requirements, including Know Your Customer (KYC) and Know Your Business (KYB) checks. Later laws, summarized in FinCEN's history of AML laws, extended these obligations to more types of institutions, including fintechs.

The cost of falling short is real. Global penalties for AML, KYC, sanctions, and customer due diligence failures totaled $3.8 billion in 2025, according to Fenergo's annual enforcement report.

The good news: better AML monitoring speeds up onboarding and strengthens fraud prevention at the same time.

Check out Alloy's glossary of compliance terms

The biggest AML compliance challenges

 

Banks, credit unions, and fintechs struggle with several key issues in AML transaction monitoring:

  • A high number of false positive alerts
  • Resource constraints from a high volume of manual reviews
  • A lack of quality data across all client interactions and products
  • Legacy systems that cannot adapt to new data sources or workflows
  • Shifts in regulatory requirements and supervisory expectations
  • Difficulty detecting complex money laundering schemes as fraud techniques evolve
  • Preparing and submitting SARs in a timely manner
  • Staying compliant when offering new products or entering new markets

The volume and complexity of financial transactions make suspicious activity hard to identify. Although core AML legislation is long established, institutions still need to monitor the adequacy of their controls as financial crime evolves, regulatory priorities shift, and new legislation changes existing requirements.

Challenge #1: False positive alerts, manual reviews, and resource constraints

False positive alerts during onboarding

There are many scenarios that cause a false positive alert during the onboarding process. To name a few:

  • A consumer could innocently provide inaccurate information by mistyping their address or other personally identifiable information (PII).
  • A consumer might share the same name or similar PII with a totally different individual who is on a sanctions watchlist.
  • A business trying to open an account could have a complex ownership structure.
  • If the data source or AML monitoring system has not been updated, it might be contributing to false positive alerts because it does not have the latest information needed to accurately identify suspicious activity.
  • The lack of real-time decisioning at account opening and false positives lead to manual review. 

When these examples result in false positive alerts, and subsequently require manual reviews, the onboarding process slows down, and customers have to wait longer to access the services they need.

False positive alerts after onboarding

AML monitoring doesn’t end after onboarding. AML regulations mandate that banks and fintechs have transaction monitoring processes in place to identify and report suspicious transactions. Large cash deposits, large withdrawals, or rapid transfers between accounts can trigger false positive alerts. However, many customers have legitimate reasons to engage in this behavior:

  • A consumer could be making a large purchase, like buying a new car or a home.
  • A small, local business, like a contractor or a roofer, could occasionally receive a large cash payment from a client. (Home repairs and renovations are expensive! And not all clients like to write checks.)
  • A consumer or business might want to open an additional account with an initial deposit that deviates from the baseline of normal customer behavior.

An AML system might identify these behaviors as potential signs of money laundering because the algorithms do not factor in the nature of the purchase or the account holder’s industry. Instead, the system flags any deposit, withdrawal, or transfer over a certain amount, even if the account has not seen any previous suspicious activity, and the transactions are completely legitimate.

Unfortunately, once an alert is generated, further investigation is required. This results in more unnecessary manual reviews, which waste time, money, and resources.

Manual reviews and resource constraints

Many alerts require a human decision, and AML teams can’t scale headcount on demand. Even a routine false positive can mean an analyst pulls customer records, compares the activity to expected behavior, documents the rationale, and closes the case, only to confirm nothing was wrong.

That adds up fast. In a March 2026 report by Liminal, 53% of banks said they spend an hour or more on each AML alert, and 37% manually review more than 40% of their alerts.

The math gets harder as institutions grow. More customers, products, and transactions typically mean more alerts, while compliance headcount and budgets often stay flat. Teams are left choosing between hiring to keep up, accepting slower reviews, or letting backlogs build. Each option has a cost. Slower reviews mean customers wait longer for account access or funds, and backlogs can put regulatory timelines at risk.

How do false positive alerts create friction?

If a false positive alert occurs during KYB and KYC checks, a bank or fintech might need to request additional documentation or information from customers to verify their identity and delay access to the new account. In either case, it is inconvenient and could result in higher customer churn or the loss of customers’ trust.

For transaction monitoring teams, they cause “alert fatigue.” A smoke detector that goes off every time someone makes toast eventually loses its urgency. AML alerts work the same way. When most alerts turn out to be nothing, analysts start to expect nothing, and the one alert that matters gets a quick glance like the dozens before it.

That’s alert fatigue, and it’s a risk in its own right. Repetitive, low-value reviews, often spread across multiple systems just to verify one entity, lead to human error and burnout. The result is a team that’s busier than ever and less able to catch what matters.

The goal isn’t fewer alerts for their own sake. It’s fewer wasted alerts, so analysts can spend their time on the cases that deserve it.

Challenge #2: Data gaps across the customer lifecycle

How does the quality and timeliness of data affect AML monitoring?

Financial institutions and fintechs need different data at different stages of the customer lifecycle:

  • Onboarding and sanctions screening need current data. KYC and KYB checks initially occur when the account opens, while sanctions screening happens both at onboarding and for ongoing payments. Stale data can let a sanctioned person open an account or move money.
  • Transaction monitoring needs complete history. Monitoring looks for patterns that build over days or weeks, like structuring or funneling. It needs every account, product, channel a customer uses, and what the signals from onboarding.
  • Bad data skews risk ratings. Outdated records can rate a high-risk customer as low-risk, or the reverse. This means either risk can be overlooked or unnecessary manual reviews.

When AML monitoring can’t see the full picture, suspicious activity gets caught too late, which can lead to a late SAR filing.

Why is it difficult for banks and fintechs to implement more integrations to different data sources?

Every stage of AML compliance runs on different data. KYC needs identity data, while KYB also needs business registries and ownership records. Sanctions screening requires current watchlists, while fraud prevention incorporates device and location signals. Each one usually means another data vendor and another integration to manage in-house.

Vendor sprawl increases when a business launches a new product or enters a market, bringing new requirements that legacy or in-house systems cannot keep pace with.

Where that data ends up matters just as much. When risk signals are spread across separate tools, it never forms one holistic view of the customer. Transaction monitoring can’t see what onboarding learned, and fraud and compliance teams work from different records.

Monitoring works best when it knows who the customer is, such as what kind of business they run, where they’re located, and how risky they are. When onboarding, fraud, and transaction data sit in one place, monitoring can use that context to tell normal activity from suspicious activity. That means fewer false alarms and faster reviews.

Challenge #3: Keeping up

Why do banks and fintechs struggle to keep up with AML requirements?

Keeping up is as much a workload problem as a regulatory one. In a March 2026 survey by Liminal, 53% of banks said they spend an hour or more on each AML alert, and 37% review more than 40% of their alerts by hand.

Banks, credit unions, and fintechs need to re-evaluate their systems regularly as legal requirements, supervisory expectations, law enforcement priorities, and available technology change. Regulators adjust their focus as transactions cross borders and technology grows more sophisticated.

Technological innovation may also impact industry standards and supervisory expectations. More institutions now use AI in compliance and fraud prevention, and regulators are encouraging it. FinCEN's April 2026 proposed rule on AML programs would have FinCEN consider whether a bank uses innovative tools such as artificial intelligence when it weighs enforcement and supervisory action. The proposal acknowledges these tools without mandating them, and it's still a proposal as of early October 2026. Examiners expect institutions to validate models, explain how alerts are generated and resolved, keep a complete audit trail, and keep people accountable for final decisions.

So, it becomes easier to see why banks and fintechs must always keep their eyes on new trends and reevaluate their compliance programs, and why they end up allocating so much time and so many resources to ensure they stay compliant.

Challenge #4: The increase in sophisticated fraud threats

Where do fraud and AML compliance intersect?

Fraud and AML compliance may intersect when funds that are illicitly gained from fraudulent activities are laundered to make them appear legal. 

The fact of the matter is, as new technology emerges to help banks and fintechs fight fraud, fraudsters are also benefiting from better tech. Just as banks, credit unions, and fintechs use AI to monitor for money laundering, fraudsters use automation and generative AI to build convincing fake identities, documents, and messages. In Alloy's 2026 State of Fraud Report, 91% of decision-makers said they see more financial crimes committed with AI. FinCEN has also warned that criminals use deepfake media to get past identity verification and open accounts that launder fraud proceeds.

The emergence of these sophisticated fraud schemes makes it more challenging to distinguish between fraudulent and legitimate activities, which creates resource strains on the fraud and compliance teams working to identify potential fraud types and criminal activity in alerts.

Money laundering and fraud share common indicators like sudden, large transactions or unusual behavioral patterns, but distinguishing between the two is critical to ensure that banks and fintechs respond appropriately. If an AML violation is misidentified as an instance of fraud, it leads to inaccurate regulatory reporting. This, in turn, could lead to increased regulatory scrutiny, fines, or penalties.

Read Alloy's Annual State of Fraud Benchmark Report

How can banks and fintechs improve AML monitoring?

To strike a balance between regulatory compliance, fraud risk management, and customer expectations, banks and fintechs need to implement more sophisticated AML monitoring solutions. They need systems and AML tools with automated workflows that:

  • Provide more high-quality data that accurately verifies customer identities
  • Breakdown data silos that impede efficient integration, sharing, and data analysis
  • Build holistic customer risk profiles across the entire customer lifecycle
  • Make rules easy to change and safe to test. 
  • Speed up SAR and CTR filing

When data is equally shared across compliance and fraud prevention teams, they can use the additional data points to enhance overall risk visibility and the identification of suspicious behavior. As a result, both teams operate as a fraud and anti-money laundering (FRAML) hub, collaborating with one another and optimizing their strategies and processes in tandem, rather than continuing to work within the confines of their own data silos.

Financial crime typologies monitoring should catch

Required monitoring looks beyond single transactions to patterns that show how illicit funds move. The Financial Crimes Enforcement Network’s (FinCEN’s) AML/CFT National Priorities identify the threats it considers most serious to the U.S. financial system, including fraud, cybercrime, terrorist financing, and human trafficking, and point to money mules and funnel accounts as ways proceeds get laundered. Patterns monitoring should catch include: 

  • Structuring: breaking up deposits or transfers to stay under reporting thresholds
  • Circulation of funds: money flows through an account, with outflows roughly matching inflows over a short window
  • Money mule activity: accounts used to receive and quickly forward funds
  • Funneling: deposits from many accounts or sources feeding a single account and move out quickly

What examiners expect from model risk management in AML monitoring

In April 2026, the OCC, Federal Reserve, and FDIC replaced SR 11-7 with a principles-based model risk management framework that scales to each model's risk. It doesn't yet cover generative or agentic AI; however, banks should still apply broader risk management and governance practices to those tools. In its April 2026 proposal, FinCEN also said it shares industry concerns about applying model risk principles to AML programs and plans to work with bank supervisors on them, so expectations are still taking shape. 

Generally, examiners are looking for the following. 

  • Model validation. Document what the model does, test it before launch, and revalidate on a schedule.
  • Explainability. Be able to show why an alert was raised and why it was closed or escalated. Reason codes and a clear alert history make that possible.
  • Audit trail. Keep a record of data inputs, scores, analyst actions, and approvals for every alert and case.
  • Human oversight. Analysts and compliance leaders stay accountable for final decisions, including suspicious activity report (SAR) filings.
  • Back-testing and tuning. Test threshold and model changes against historical data before going live, and record the results.
  • Data quality controls. Show that the data feeding monitoring is complete, accurate, and current.
  • Funneling: deposits from many accounts or sources feeding a single account and move out quickly

Signs you need to replace your legacy AML monitoring system

Legacy systems often keep working long after they stop working well. If several of these sound familiar, it’s time to upgrade your AML transaction monitoring system:

  • Alert volume keeps rising but confirmed suspicious activity does not. Analysts spend most of their time clearing false positives.
  • Adding a new data source takes months. Integrations are rigid, expensive, or need custom engineering.
  • Rules are hard to change or not able to be changed at all. Tuning thresholds requires vendor tickets or IT projects.
  • You can’t test changes before going live. There is no back-testing or simulation, so every change carries risk.
  • SAR preparation is manual. Analysts copy data between systems to build narratives and file reports.
  • Compliance and fraud teams work from different data. Alerts lack the context of onboarding, device, and behavioral signals.
  • Audit and exam preparation is a scramble. Evidence for why decisions were made is scattered across systems.

Alloy works with your existing data and workflows and makes your legacy systems stronger, so you can prove results on real alerts before you expand.

What is an identity and fraud prevention platform?

An identity and fraud prevention platform is an end-to-end platform that helps manage compliance, identity, fraud, and credit risks throughout the customer lifecycle. Some teams call this an Identity Risk Solution.

It helps banks and fintechs meet AML, KYB, and KYC standards and mitigate fraud threats by:

  • Integrating multiple data sources to improve the quality of data and provide real-time data that helps build more accurate customer profiles
  • Using advanced analytics and machine learning algorithms to improve the accuracy of alerts and their prioritization while reducing the number of false positives
  • Automating regulatory requirements and aggregating suspicious activity reports (SARs) that can be e-filed directly with the Financial Crimes Enforcement Network (FinCEN)

It also utilizes adaptive risk scoring models to:

  • Better evaluate customer risk levels
  • Adjust risk thresholds based on evolving customer behavior
  • Change risk profiles and reduce the chances of overlooking new risks or more complex schemes
  • Customize control processes for individual accounts
  • Apply risk-based authentication across the customer lifecycle, so low-risk customers move through seamlessly, and high-risk activity triggers step-up verification

Explore all aspects of an identity and fraud prevention platform

What are the benefits of an identity and fraud prevention platform?

An identity and fraud prevention platform has a multitude of benefits for banks and fintechs:

  • It centralizes data, so compliance and fraud teams have a clearer view of customers’ activity history.
  • It combines the data used during onboarding with the data acquired during ongoing monitoring to provide a more robust view of every customer and establish a better baseline of good customer behavior.
  • It runs recurring, automated checks against sanctions watchlists.
  • It provides the flexibility that’s needed to comply with new regulations as they change and evolve, so banks and fintechs can quickly adapt and incorporate new data sources into their workflow.
  • It helps identify more sophisticated instances of suspicious activity, so it’s easier to comply with current regulations.
  • It helps move banks and fintechs towards a perpetual KYC approach instead of an irregular, periodic KYC approach.

In summary, an identity and fraud prevention platform keeps pace with both compliance regulations and fraud prevention from the very beginning, while reducing customer friction. And those protections don't stop once customers are onboarded, they continue throughout the customer lifecycle. In turn, suspicious activity is detected more efficiently and effectively. An identity and fraud prevention platform also enables compliance and fraud teams to collaborate and conduct joint application reviews when necessary, then work together to resolve issues.

How Alloy can help

Alloy is the AI-powered identity and fraud prevention platform that accelerates onboarding, stops fraud, and scales compliance across the customer lifecycle. It connects banks, credit unions, and fintechs to 300+ data solutions through one integration, so monitoring runs on the data quality it depends on.

With Alloy Transaction Monitoring, teams can:

  • Run out-of-the-box and custom AML monitoring rules in real time
  • Back-test rule and threshold changes against historical data before going live
  • Detect typologies such as structuring, circulation of funds, money mule activity, and funneling
  • Bring onboarding, fraud, and transaction data into one view of the customer, so compliance and fraud teams work from the same facts

Alloy's AI Assistant, part of Alloy's Actionable AI suite, handles the AML investigation work that slows analysts down. AI Analysis explains why an alert is in its current state and what should happen next, with the thresholds and data behind it. Suspicious Pattern Analysis looks across past alerts, investigations, and event history for patterns a single alert won't show. Analysts review and decide in the same platform, so every decision stays in one record.

One integration gives you access to Alloy's 300+ data solutions, and a no-code rules editor lets your risk team change monitoring rules without engineering support. Step-up verification runs automatically when risk calls for it, and the Testing Suite shows projected outcomes for new data solutions and rule changes before you go live. With Alloy, banks, credit unions, and fintechs scale AML monitoring at the pace of growth.

Alloy helps banks, fintechs, and credit unions turn compliance risk management into a competitive advantage.

More resources

Frequently asked questions (FAQs)

The most common weaknesses are high false positive rates, heavy manual review workloads, poor or outdated data, rigid legacy systems that are hard to integrate or tune, limited ability to detect complex typologies, and weak separation between compliance and fraud data. Together, they slow onboarding, raise costs, and increase the risk of missing real suspicious activity.

Look for rising alert volume without more confirmed cases, long timelines to add data sources, rules that are hard to change or test, manual SAR preparation, siloed compliance and fraud data, and difficulty producing evidence for examiners. 

Yes. Regulators have encouraged responsible innovation in AML programs since the agencies' December 2018 joint statement on innovation. FinCEN's April 2026 proposed rule would have FinCEN consider whether a bank uses innovative tools such as artificial intelligence when weighing enforcement and supervisory action. Institutions are expected to validate models, explain alerts, keep audit trails, and keep people accountable for decisions.

AI Analysis in Alloy’s AI Assistant explains why an alert is in its current state and what should happen next, with the thresholds and data behind it. Analysts review and decide in the same platform, so every decision stays in one record.

Connect higher-quality data, back-test rule changes before launch, and use risk-based alert prioritization. Continuous tuning matters as customer behavior changes. Alloy evaluates each transaction against the customer's full identity and activity profile, including KYC data and onboarding history, so alerts reflect context. 

Funds gained through fraud are often laundered, and both disciplines look for sudden large transactions and unusual behavior. Sharing data between fraud and compliance teams (a FRAML approach) gives both a fuller view of risk. Alloy brings identity, onboarding, fraud, and transaction data into one customer view with shared case management, so both teams work from the same facts.


See what you’re missing

First, we’ll learn about your needs, answer your questions, and then see how Alloy can help.
Back