Beyond cost savings: AI as a force multiplier for financial crime operations

AI should change what a team can accomplish—not simply how quickly it completes existing tasks.

For financial institutions evaluating AI in financial crime operations, the ambition should extend beyond processing alerts faster. The opportunity is to reconsider how work is performed, where experienced analysts devote their attention, and what additional capability the institution could gain.

What McKinsey argues

In Growth favors the bold: AI as force multiplier, McKinsey examines seven myths that constrain AI-driven commercial growth. The authors group them into three categories: myths that limit ambition, preserve old ways of working, and slow down value capture. Their central argument is that organizations need to redesign how commercial decisions are made, rather than simply deploy more technology.

The article challenges the idea that AI is primarily a productivity tool. It emphasizes effectiveness alongside efficiency, business and technology leadership working together, and investment in new skills, ways of working, and governance. It also challenges the assumptions that companies need perfect data before starting or must wait a long time to see value. In its commercial examples, McKinsey distinguishes improvements in leading indicators within weeks from substantial economic value within three to six months.

The article concerns commercial growth—not AML operations. The application to financial crime below is Silk River’s interpretation, not a claim that McKinsey has evaluated AML Resolve.

The Silk River perspective: make the operation more capable

Reducing manual effort is valuable. But it should be the beginning of the conversation, not its limit.

An institution evaluating AI should also ask what its financial crime team could accomplish with the capacity released from routine alert handling. Could experienced analysts spend more time on complex investigations? Could the institution devote additional attention to enhanced due diligence? Could it reduce an existing backlog or accommodate a growing workload without a proportionate increase in staffing?

These are outcomes to evaluate, not benefits to assume. They nevertheless establish a more useful ambition than making individual tasks faster.

The objective should be a more capable financial crime operation: one that uses analyst judgment where it matters most and measures whether the redesigned workflow improves both productivity and the quality of the work.

Keep the systems. Rethink the work.

Workflow redesign does not necessarily require replacing existing systems.

An overlay describes how software integrates. Workflow redesign describes how the work gets done. An institution can retain its screening and case-management environment while changing how evidence is assembled, cases are prepared, and decisions are routed.

AML Resolve operates within that existing environment. It gathers approved context, assembles evidence, prepares rationales, and routes alerts toward customer-approved auto-clear eligibility, analyst-ready review, or escalation for higher-risk or ambiguous cases.

That is a different proposition from giving every analyst a faster way to write the same narrative. The aim is to change the distribution of work between automation and human judgment, with the institution defining the boundaries.

Keep the systems you rely on. Rethink the work performed around them.

Put released capacity to work

A financial crime leader should decide in advance how any capacity gains will be used.

For one institution, the priority may be reducing overtime or external support costs. For another, it may be avoiding additional hiring as volumes grow. For a third, the greatest value may come from redirecting experienced analysts toward complex cases and enhanced due diligence.

These benefits should be measured separately. Hours released are not automatically cash savings, and the same hours should not be counted both as eliminated expense and as capacity redeployed to additional work.

The evaluation should therefore consider more than handling time. It should also examine the amount and quality of work completed, evidence completeness, rework, backlog reduction, and the capacity available for higher-priority investigations.

Measure the opportunity, then test it

Silk River’s AML Performance Lab processes a meaningful sample of an institution’s historical alert workload before an operational pilot begins. It develops a customer-specific projection of potential changes in manual effort, resolution paths, analyst capacity, and unit cost.

A bounded AML Resolve pilot can be configured and launched in weeks, subject to the institution’s approvals and data readiness. The pilot is designed to run for approximately four weeks and to validate or refine the Lab’s projections under operating conditions.

For this broader evaluation, the question is not only whether an alert takes less time to resolve. It is whether the institution can use the resulting capacity to accomplish more of the work it considers important, while maintaining its required controls.

Our companion article, “AI adoption is rising. ROI is not. AML resolution is a practical place to prove value,” explains the benchmarking and pilot approach in greater detail.

The goal: more than a lower cost per alert

The strongest case for AI in financial crime operations should connect efficiency to a clear operational purpose.

Start with the work that consumes capacity. Determine which activities can be automated or prepared for review, where human judgment remains essential, and how released capacity would be used. Then measure the operational improvements and the financial implications separately.