AI Centers of Excellence

AI Center of Excellence Leader, PwC Switzerland

Email

Artificial intelligence (AI) is transforming the financial services industry, with its potential to offer banks greater operational efficiency, improved risk management and enhanced customer experience. The challenge for banks is how to scale AI capabilities efficiently, whilst accommodating the diverse needs and priorities of different banking functions.

Take retail banking, where AI-powered chatbots can provide customers with better support. In wealth management, AI enables banks to offer hyper-personalised recommendations to clients. And in corporate banking, AI can improve credit risk assessments.

To fully benefit from AI’s potential across the diverse use cases of different business functions, banks need a strategy that incorporates centralised control with divisional flexibility. In this blog, I'll explore how embedding divisional AI Centers of Excellence (CoEs) within a network of federated CoEs, combined with a scalable GenAI factory approach, can help financial services firms maximise the value of AI adoption.

What are divisional AI CoEs?

Divisional AI Centers of Excellence are specialised units within specific business areas, such as asset management, that focus on developing and implementing AI solutions tailored to the division’s unique needs. These CoEs serve as hubs of expertise, fostering innovation and ensuring that AI initiatives align with divisional goals and challenges.

For instance, in asset management, a divisional AI CoE might focus on developing AI models for portfolio optimisation, risk assessment, or client personalisation. By concentrating expertise and resources, these CoEs can drive targeted AI adoption that delivers tangible business value.

Amplifying impact through federated CoEs

While divisional CoEs are powerful in their own right, connecting them into a network of federated CoEs can enhance their impact across the organization. Taking a federated approach allows for:

Sharing of knowledge sharing and best practices

Standardisation of AI governance and ethical frameworks

Efficient resource allocation and reduction in duplicate efforts

Cross-divisional collaboration on AI initiatives, minimising silos

In a financial services context, this could mean that innovations in AI-driven customer service developed by the retail banking CoE could be quickly adapted and implemented by the asset management division, accelerating overall AI adoption.

The scalable GenAI factory approach

A scalable GenAI factory acts as a centralised hub for developing and maintaining foundational AI capabilities. This approach involves:

Creating a value hypothesis for GenAI applications
Prioritising key use cases across the organisation
Developing use cases patterns or foundational capabilities to drive scale
Selecting foundational GenAI tools
Defining solutions that maximise existing value

By centralising these core capabilities, organisations can rapidly develop and deploy AI solutions across various divisions, ensuring consistency and efficiency in their AI initiatives.

Crucially, this centralised-decentralised setup must provide more than just a central library of foundational capabilities. It should also establish:

A comprehensive governance framework that ensures all AI initiatives adhere to regulatory standards and ethical guidelines
Robust quality control processes to maintain the integrity and reliability of AI models used across the organisation
Standardised evaluation methodologies to assess the performance and impact of AI solutions consistently
Clear protocols for data privacy and security that comply with relevant regulations (such as GDPR, CCPA)
A centralised risk management approach to identify, assess, and mitigate AI-related risks

This governance-centric approach ensures that as financial institutions scale their AI capabilities, they do so responsibly and in compliance with regulatory requirements. It provides a balance between innovation and control, allowing divisions to leverage AI effectively while maintaining enterprise-wide standards and safeguards.

Where the real power lies

The real power lies in combining the federated CoE network with the GenAI factory approach. By doing so, banks will be able to :

Take advantage of centralised AI capabilities while maintaining divisional expertise

Rapidly scale successful AI solutions across the organisation

Ensure consistent governance and ethical standards

Foster innovation through cross-pollination of ideas

For financial institutions, some ways that foundational capabilities could be reused across multiple divisions include:

Wealth management

AI-driven client profiling and risk assessment models developed here could be adapted for use in retail banking for personalised product recommendations.

Corporate banking

AI systems for fraud detection and anti-money laundering could be scaled across retail banking and wealth management divisions.

Investment banking

Natural Language Processing (NLP) models for analysing market sentiment and financial reports could be repurposed in corporate banking for credit risk assessment.

Asset management

Portfolio optimisation algorithms could be adapted for use in wealth management for client portfolio construction.

How to get started?

To successfully implement an integrated federated CoE network with a GenAI factory, financial services firms should consider the following:

Establishing clear governance structures that define the roles and responsibilities of divisional CoEs, the federated network, and the central GenAI factory.

Investing in data infrastructure that allows for secure and efficient data sharing across divisions while maintaining regulatory compliance.

Developing a talent strategy that balances centralised AI expertise with domain-specific knowledge in divisional CoEs.

Implementing continuous learning and adaptation processes to keep pace with the rapidly evolving AI landscape.

Fostering a culture of collaboration and innovation that encourages cross-divisional AI initiatives.

Final thoughts

As financial institutions continue their digital transformation journey, the ability to scale AI capabilities effectively will be a key differentiator. By embedding divisional AI CoEs within a federated network and combining this with a scalable GenAI factory approach, banks can create a powerful ecosystem for AI innovation and adoption.

This integrated strategy balances centralised efficiency with divisional expertise, ensuring that AI initiatives are both scalable and tailored to specific business needs. And with AI increasingly becoming business as usual, those that master embedding this technology into their operating models will be best positioned to lead in their respective markets.

If you’d like to discuss embedding GenAI Centres of Excellence or implementing a scalable GenAI factory, please get in touch.

Contact us