Skip to Content

Activating Gen AI at scale to transform financial services: Insights and best practices from Capgemini

Ashvin Parmar
06 November 2024

Generative AI (Gen AI) is transforming nearly every industry, and financial services is no exception. As banks and insurance companies navigate an increasingly competitive and complex business landscape, harnessing the power of Gen AI can unlock significant business value. While nearly every financial institution has begun experimenting with and piloting this transformative technology, only a few have successfully scaled their Gen AI models for widespread production.

AI agents promise smoother automation and enhanced productivity

Source: Capgemini Research Institute, Generative AI executive survey, May–June 2024, N = 1,031 organizations who are at least exploring Generative AI capabilities

In this article, we explore how financial services organizations can effectively implement and scale Gen AI to transform their operations and drive innovation.

Scaling and operationalizing Gen AI

In recent months, many financial services firms have shifted from experimenting with Gen AI to scaling up and operationalizing use cases. This transition is fueled by the recognition of Gen AI’s transformative potential, particularly in enhancing business processes and customer experiences. To maximize the value of Gen AI investments, financial institutions should focus on several key activities:

1) Identify high-impact use cases

A crucial first step in activating Gen AI is pinpointing use cases that have a high impact on business goals and can be developed within a reasonable timeframe. It’s important to avoid burdening these initiatives with unrealistic expectations that cannot be implemented promptly.

In financial services, many organizations are focusing their attention on the following key areas:

  • Personalized customer experiences: Gen AI can tailor product recommendations, generate targeted content, and automate communication and marketing campaigns, boosting customer satisfaction and loyalty.
  • Enhanced risk management: Gen AI can significantly improve processes like fraud detection, credit risk assessment, and anti-money laundering (AML) compliance, increasing operational efficiency and accuracy.
  • Streamlined operations: Automating repetitive tasks such as report generation, data analysis, and customer service interactions allows human capital to focus on higher-value activities.
  • Expert-in-the-loop: Gen AI-powered contact center co-pilots can assist customer service representatives by providing real-time guidance on products, services, and support questions.
  • Tailored technology integration: Instead of relying solely on Gen AI, organizations can leverage a combination of technologies, including other AI models and Robotic Process Automation (RPA).
  • Governance and trust: Creating robust governance frameworks helps organizations maximize AI’s potential while mitigating risks. Transparent, consistent, and accountable AI implementation promotes stakeholder collaboration, ensuring that strategies prioritize risk management, ethical practices, and alignment with organizational values.
  • Establish a value framework: Create a system for business users to experiment with, measure, and benchmark Gen AI applications. This enables the quantification and justification of high-value use cases.

High Impact use case selection process

2) High Impact use case selection process

Financial Services firms have been heavily investing in cloud infrastructure to address legacy technical debt and democratize access to data. Augmenting these platforms with tools like Capgemini’s new RAISE (Reliable AI Solution Engineering) accelerator enables faster deployment and management of Gen AI solutions.

3) Focus on data quality and management

Gen AI models rely heavily on high-quality data. Financial institutions are investing in several critical areas to ensure optimal AI performance. Data cleansing and standardization are vital to maintain accuracy, completeness, and consistency across various sources. Proper data labeling and annotation guide AI models to learn desired patterns effectively. Additionally, establishing clear data governance frameworks is crucial. These frameworks define ownership, access control, and security protocols to protect sensitive information, thereby enhancing the reliability and integrity of Gen AI models.

Real-world implementations

Financial services organizations are leveraging Gen AI in diverse ways, often employing a blend of automated and human-augmented processes.

Adoption of generative AI has grown across functions

Source: Capgemini Research Institute, Generative AI executive survey, April 2023, N = 800 organizations; Generative AI executive survey, May–June 2024, N = 1,031 organizations who are at least exploring Generative AI capabilities

Here are several examples we’re working on now with clients:

Legacy modernization

Gen AI excels in tasks like code documentation, code generation, and test case creation. However, maximizing code reuse while refactoring legacy systems is complex and typically involves the following steps:

  1. Documentation: Explaining the purpose and functionality of the code to help developers understand the rationale behind specific changes.
  2. Discovering code relationships: Identifying and mapping dependencies between different code elements to streamline the refactoring process.
  3. Analysis and design for target architecture: Deciding how the code should be structured, which patterns to apply, and how components will interact in the new architecture.
  4. Code generation: Analyzing the code to identify areas for improvement, such as eliminating code smells (e.g., duplicated code, long methods, complex logic, or inconsistent naming conventions).
  5. Generate test cases and test data: Automatically generating scenarios and input data for testing purposes, ensuring that refactoring does not inadvertently introduce bugs or regressions.

Business operations

In banking and insurance operations, back-office processes are being examined for automation opportunities using Gen AI. Examples include dynamic Q&A systems and automated document creation. The principle here is to automate tasks where feasible and augment human effort where automation is not possible, thereby enhancing and transforming operations.

Business processes

For business-critical functions like legal, finance, and underwriting, financial services firms are using Gen AI to augment human capabilities in key areas:

  • Surfacing answers from data: Searching large sets of documentation to support more informed decision-making.
  • Learning from historical data: Utilizing historical data from service desks and other sources to improve manual adjudication processes.
  • Summarizing information: Providing starting points or drafts for documentation, streamlining the creation process.
  • Reimagining business processes: Combining Gen AI with existing tools and AI/ML models to enhance and optimize business operations.

Precautions for deploying Gen AI

When implementing Gen AI solutions, financial services companies must consider several critical areas before developing or deploying any models:

  • Data privacy and security: Financial data is highly sensitive, making data privacy and security paramount when using AI models for generation or analysis. Robust encryption, strict compliance with regulations (such as GDPR or CCPA), and strong defenses against cyber threats are essential to protect against privacy breaches and security vulnerabilities.
  • Biases (intentional and unintentional): Gen AI models can inherit biases from the data they are trained on, potentially leading to unfair outcomes in financial decision-making. This can result in discriminatory practices or skewed recommendations, exposing institutions to regulatory scrutiny and reputational damage. Financial firms must rigorously evaluate model outcomes during the training phase to identify and address any biases. Additionally, augmenting in-house data with third-party data can help overcome inherent limitations and provide a more balanced perspective.
  • Regulatory compliance: Financial Institutions operate within stringent regulatory frameworks (e.g., Basel III, Dodd-Frank, etc.). The introduction of AI models requires compliance with these regulations, which often do not explicitly address AI technologies, leading to uncertainties. Careful interpretation of existing regulations is needed to ensure that AI applications align with compliance requirements.
  • Interpretability and transparency: AI models, particularly complex generative models, often function as black boxes, making it difficult to understand how they arrive at decisions or generate outputs. This lack of transparency poses challenges for financial institutions that must justify their decisions and ensure accountability. To address this, banks and insurers should carefully document input data and its relevance to the model’s decision-making process, creating a consistent audit trail that supports transparency and facilitates compliance reviews.

Best practices for winning with Gen AI

Successful Gen AI adoption requires a well-defined strategy, persistence, and a willingness to learn from previous experiences. Here are some best practices to guide financial services firms in their AI initiatives:

  • Educate and align: Educating stakeholders at every level—from board members to developers—is critical for Gen AI success. For instance, our own Gen AI Center of Excellence (COE) hosts a bi-weekly Gen AI Hour meeting, where key stakeholders are updated on successful case studies and cross-sector use cases. This forum helps keep everyone informed, aligned, and engaged with the AI strategy.
  • Build a culture of AI readiness: Invest in employee training programs to build awareness, understanding, and comfort with AI technologies. Cultivating an organization-wide mindset that embraces AI is essential for successful Gen AI adoption.
  • Establish trust and governance: Address ethical considerations related to bias, fairness, and data privacy by setting clear ethical guidelines and committing to responsible AI development practices, such as those outlined in Capgemini’s Code of Ethics for AI. Strong governance frameworks will help build trust and promote ethical AI use across the organization.
  • Prioritize continuous improvement: Gen AI is an evolving field. Establish a culture of continuous learning and improvement by regularly monitoring model performance, gathering feedback, and iterating on models.
  • Partner and co-invest: Collaborate with partners to develop joint solutions, educate partner sales organizations about your Gen AI go-to-market strategies, and sell solutions together.

Capgemini’s Comprehensive AI framework

Ready, set, go!

Gen AI is reshaping the financial services industry, providing banks and insurers with powerful tools to drive innovation, improve customer experiences, and streamline operations. While many institutions have started experimenting with Gen AI, the real challenge lies in effectively scaling these models for widespread use.

Success in this space requires a strategic approach, focusing on high-impact use cases, leveraging cloud infrastructure, and prioritizing data quality and governance. Real-world implementations show that Gen AI can modernize legacy systems, optimize business operations, and augment key business processes. However, it’s equally important to address concerns around data privacy, bias, regulatory compliance, and model transparency to ensure ethical and responsible AI adoption.

By educating stakeholders, fostering a culture of AI readiness, building robust governance frameworks, and continuously improving AI models, financial services firms can unlock the full potential of Gen AI. Partnering with the right allies will further amplify success, enabling organizations to navigate the complexities of this transformative technology and achieve a competitive edge in the industry.

As a recognized “Leader” in The Forrester Wave: AI Services, Q2 2024 report, Capgemini is uniquely positioned to help your organization activate and scale Gen AI. Contact us today to start your AI journey.

Special thanks to: Clement de Balby de Vernon, Ryan Toa, Tom Nicholson

Meet our experts

Ashvin Parmar

Global Head of FS Generative AI CoE

Lars Boeing

Expert in AI in FS, Capital Finance, Digital Transformation
Principal, Invent Financial Services at Capgemini Invent North America Lars Boeing leads the Insurance team for Capgemini Invent North America, focusing on supporting carriers in their digital transformation and in becoming inventive insurers.

Rajesh Iyer

Global Head of AI and ML, Financial Services
Rajesh is the Global Head of AI and ML for Financial Services. He has almost three decades of of experience in the Financial Services Industry, working with Fortune/Global 500 clients seeking to maximize the value of investments in their Enterprise Data and AI programs.

Shankar Ramanathan

Senior Director, AI & Machine Learning, Financial Services Insights & Data

    Expert perspective