Skip to content
Industry focus

Financial Services

AI transformation across banking, financial services and insurance. Our focus connects practical use cases with the operating decisions behind them: how work changes, where people stay involved and what evidence supports wider adoption.

Transformation Services

Connect industry experience with your next decision through advisory, executive workshops, fractional leadership and executive search.

See Transformation Services

Financial services research

Evidence for the next decision.

Five primary sources on AI adoption, governance and operating risk. Each connects published evidence with a question to bring into your evaluation.

01

Adoption and governance

Where AI is used, and who remains accountable

Bank of England and FCA · November 21, 2024

Read the original publication ↗

What the source establishes

The joint survey examines AI use, third-party dependencies, automated decisions and governance across UK financial services firms.

The executive question

Who owns the decision when AI supports or takes an action?

Institute interpretation

Map each proposed use case to its data, provider dependencies, accountable owner and approval points. The survey describes its participating firms; it does not establish the return on your project.

02

Broker-dealer operations

The tool changes. Existing obligations still apply.

FINRA · Regulatory Notice 24-09 · June 27, 2024

Read the original publication ↗

What the source establishes

FINRA reminds member firms that its technology-neutral rules continue to apply when they use generative AI. The notice does not create new requirements.

The executive question

How will supervision and review work for this AI-assisted process?

Institute interpretation

Evaluate the full workflow, including communications, records and the person responsible for review. A successful demonstration is only one part of the operating decision.

03

Enterprise deployment

Separate the opportunity from the operating risk

U.S. Department of the Treasury · December 2024

Read the original publication ↗

What the source establishes

Treasury’s report summarizes responses to its request for information on AI uses, opportunities and risks, including data quality, explainability and third-party reliance.

The executive question

What evidence would justify moving this use case into production?

Institute interpretation

Set the expected outcome alongside the data, controls and human effort required to achieve it. The report synthesizes stakeholder views; it is not a performance benchmark.

04

Insurance

Build governance around decisions that affect customers

NAIC · Model Bulletin · December 4, 2023

Read the original publication ↗

What the source establishes

The model bulletin sets out expectations for insurers’ AI governance, risk management and documentation. It is a model for state regulators, not a uniform national rule.

The executive question

How will the insurer test and document the impact of AI-supported decisions?

Institute interpretation

Bring underwriting, claims, data and risk teams into the evaluation. Examine the applicable state requirements separately, and ask vendors for evidence that can support your own review.

05

Resilience and concentration

A workflow can depend on more than one AI provider

Bank of England · Financial Stability in Focus · April 9, 2025

Read the original publication ↗

What the source establishes

The Bank of England examines AI’s potential benefits and implications for financial stability, including interconnected dependencies and risks that can extend beyond one firm.

The executive question

What happens to the service if a shared model, cloud or data provider fails?

Institute interpretation

Include dependency mapping, fallback procedures and recovery ownership in the business case. Evaluate resilience alongside speed, quality and cost.

Banking: customer and operational workflows

Client onboarding, document review, servicing and internal knowledge are starting points for examining AI in banking. We focus on how these workflows move from a request to a completed action, including identity, data access, exceptions and handoffs between teams. The questions center on service quality, processing time and the controls needed to operate at scale.

Discuss a banking use case

Insurance: underwriting, claims and service

In insurance, our focus includes submission intake, underwriting support, claims documentation and policy servicing. We examine how AI can help teams assemble information and prepare work for review, alongside the judgment required to resolve exceptions. Each use case connects the customer experience with the evidence behind a decision.

Discuss an insurance use case

Financial services: research and knowledge

Research teams, advisers and operations leaders need to find information, compare sources and turn analysis into useful work. We examine knowledge retrieval, document synthesis and internal research support, with a focus on source traceability, review requirements and the boundaries of an agent’s authority.

Discuss a financial services use case

Measure performance in context

An AI workflow needs a defined starting point and a clear standard for completion. Our evaluation questions connect throughput, turnaround time, quality and cost with review effort, exception rates and adoption. The aim is to understand the whole process, including what happens when the system needs a person to intervene.

Read the methodology

Build your evaluation around the decision

The Institute’s research on platform evaluation, data foundations and workflow economics provides a starting point for industry questions. Use the Vendor Profiles to investigate product scope and the evidence you need for a particular workflow.

Bring us a research question from your organization. Describe the process, the team responsible and what you would need to see before expanding the use of AI.

Browse research