The Next Chapter in Fund Operations: AI in Practice

Smarter Reconciliations

Reconciliation is one of the clearest examples of how AI can improve outcomes in fund operations. For financial firms, the value is direct: less time spent by humans in investigating exceptions, faster resolution of breaks and greater certainty over positions and cash.

“At its foundation, reconciliation is a deterministic process that compares one dataset with another to confirm alignment,” says Nellie Ding, Global Head of Operations Digital Transformation, BNY. “That structured nature makes it well suited to AI.” At BNY, agentic AI reconciliation capabilities are helping accelerate the workflow by supporting data ingestion, cleansing, standardization and enrichment, while also reinforcing core data governance elements.

AI can then perform automated transaction matching, identifying patterns in transaction data and descriptions and recognizing broader relationships across datasets to improve accuracy and reduce manual intervention. When exceptions arise, AI can help accelerate resolution by analyzing root causes and suggesting effective next steps.

It can also add a further layer of intelligence to exception management by interpreting comments and contextual signals, assessing aging and risk, and helping prioritize the items most in need of attention. The result is a more proactive approach to managing breaks and a more responsive operating model overall.

AI-Powered Oversight & Controls

In NAV-oriented processes, AI is helping shift oversight beyond static thresholds and manual checks toward a more dynamic and precise approach. For financial firms, that can mean greater confidence in accuracy and fewer unnecessary alerts requiring review.

Within BNY’s global Fund Accounting business, AI-supported dynamic benchmarking and anomaly detection are helping identify unusual patterns earlier and prevent errors more proactively. Rather than relying solely on fixed tolerances, AI can review the day’s transactions and compare them individually with a fund’s own transaction history, market history, information on similar funds and other relevant data sources.

This creates a clearer focus on what matters most. In one example, BNY’s Fund Accounting team used dynamic fund benchmarking during NAV calculation to reduce false positives from 90% to less than 20%. That means fewer low-value alerts to investigate and greater attention on the exceptions that genuinely matter.

AI also brings greater analytical capacity. “Smaller anomalies that may previously have fallen below a manual review threshold can now be identified earlier,” says Bill McManus, Head of Global Fund Administration Product at BNY. “This helps teams detect issues sooner and reduce potential downstream risk.” By escalating only where needed, AI supports NAV processes that are more accurate, more efficient and more responsive.

Reimagining NAV Monitoring

NAV production remains at the center of the daily fund accounting cycle. It supports regulatory compliance, enables processing of investor flows, determines fund performance and provides critical visibility into next-day operations. The earlier the NAV can be produced, the better.

“Today we leverage intelligent NAV across several funds, while for other, more complex funds, we are steadily extending automation capabilities to further reduce manual intervention and accelerate NAV delivery,” says Nicole Greene, Head of Strategy and Product Transformation at BNY. “The process is complex, and control requirements remain paramount. Even so, meaningful progress is already being made toward a more automated model.”

Real estate funds provide a good example. Creating a NAV in this segment is often more complex because many prices must still be supplied manually. BNY has developed AI tools to help automate the gathering of this data and support data validation. High-quality data remains essential to accurate outcomes, so AI-enabled data management and governance play a foundational role in strengthening the reliability of the models used in NAV production.

BNY is also working to enable anomaly detection in transactions throughout the day, so that by the time the “golden hour” for NAV construction arrives, key controls are already under way. The increased availability of higher-quality, AI-processed data is an important enabler of this shift.

Another area of progress is the use of multiple models on a single task to increase automation while preserving control. For validation of financial statements where accuracy is critical, for example, an agentic AI model and a more traditional large language model can analyze documents independently. Where the outputs align, processing can continue automatically. Where they differ, the item is escalated for human review.

The Fund Operations Professional of the Future

As AI takes on more of the repetitive work that has long defined fund operations, the role of the operations professional is beginning to change with people bound to spend less time on routine processing and more time on oversight, judgment and client service.

For financial firms, the benefits are tangible. Earlier resolution of issues can reduce downstream errors across books and records, while tighter processing cycles can support faster settlement, faster NAV delivery and more efficient use of working capital. Time spent on exception management, manual research and document handling can also be reduced.

This evolution does not diminish the importance of operational expertise. It increases it. “As routine activity becomes more automated, the human role becomes more focused on supervising AI-enabled workflows, refining controls, managing true exceptions and helping clients navigate complexity with confidence,” says Monahan.

At BNY, that transition is being supported through investment in continuous learning, role-based development and leadership capability, helping teams adapt as operational work evolves. “The broader lesson is that best practice in operations is no longer just about scale,” says Ding. “It is about combining AI with deep expertise to deliver a model that is faster and more responsive to the evolving demands of the global financial services ecosystem.”

What's Next?

The next challenge is to connect AI tools across the operational lifecycle and integrate them more fully across engineering, product and operations. That evolution will also extend beyond institutional silos. 

At the same time, the direction of travel will continue to depend on retaining what matters most: standards, controls, regulatory compliance and robust risk management. At BNY, all AI tools must pass through a rigorous governance process before going live, with legal and compliance embedded throughout. Oversight extends across the full AI lifecycle, including deployment, continuous monitoring, change management, telemetry logging, lineage tracking, approval workflows, access recertification, incident analysis and periodic retesting against real operational scenarios. Transparency is also built into AI models to support auditability and stronger human supervision.

Fund operations are entering a new era — and AI is set to play a defining role in what comes next. The opportunity now is not only to improve individual processes, but to reimagine the operating model as a whole. For firms that can pair innovation with strong controls and operational expertise, the next phase of transformation is already taking shape.

For more information about our solutions, please visit our Fund & Investor Solutions website.