Tue, Dec 24 2024

Building Generation AI at Scale: Strategies for CIOs to Navigate Risk

May 13, 2024
4 Min Reads

We spoke with David Kadio-Morokro of EY on how CIOs may manage risk and develop Gen AI skills at scale in this in-depth interview.

AI

After releasing his EY report, Financial Services CIOs - building Gen AI at scale while managing risk, David Kadio-Morokro, Americas Financial Services Innovation Leader for EY, discusses the difficulties CIOs encounter in scaling Gen AI and how to get past them in an interview with FinTech Magazine.

Chief Information Officers (CIOs) must first see the benefits of enabling these new capabilities throughout the tech stack in order to revolutionize banking operations, drive cost savings, and drive efficiency before they can effectively deploy new Gen AI capabilities at scale.

The benefits of Gen AI for the whole tech stack


While it's evident that many CIOs understand the benefits of using Gen AI, David believes that implementing new capabilities might be challenging.

The speaker elucidates, saying that financial institutions persist in upholding substantial traditional technology portfolios, stressing the transfer of structured data from Front Line Units (FLU) to middle and back office.

"High-speed and unstructured data processing incremental investments are frequently disconnected from their current infrastructure, resulting in data proliferation and usage frictions."

In order to furnish banking colleagues with appropriate Gen AI tools for customer engagement, risk management, or compliance, David suggests that CIOs allocate resources towards the following key capabilities:

Processing unstructured data at scale increasing spending on technologies for data search, tagging, and classification in order to find hidden metadata in unstructured materials.

Novel techniques for storing data: increasing the amount of money spent on graph and vector databases, which are used to store metadata (such as entities, relationships, and attributes) integrated into documents and enable methodical querying and analysis of unstructured data.

LLM adjustment: putting money into tools that enable large language model (LLM) tweaking and fine-tuning in order to increase the LLMs' suitability for use cases unique to each firms.


The difficulties in developing these skills

Although there is a clear roadmap for implementing Gen AI transformation, CIOs must first overcome some obstacles in order to accomplish the three main goals mentioned above.

As a new frontier for CIOs everywhere, David notes that the agenda's relative youth "presents a major challenge" and "brings a plethora of learning curves along with it."

He goes on, "For instance, it's challenging to control LLMs' insatiable demand for data. The likelihood of unintentional exposure and spread of more varied data sets for Gen AI use cases rises with their applicability, raising the possibility of data security and privacy problems. On the other hand, a dearth of data for use cases might hinder innovation by costing opportunities.

Furthermore, a brand-new issue known as "vendor confusion" has emerged. It might be challenging to sort among the hundreds of suppliers offering models, data storage, libraries, and other resources and find the one that works best for you. Long testing sessions are a result of the disagreement about the highest-value use cases.

Gen AI: A trinity of people, processes, and technology


David argues that in order to maximize the impact of Gen AI, CIOs and IT operations must effectively balance people, procedures, and the technology itself.

David states that "cross-collaboration across all layers of the organization is required for a successful implementation of Gen AI." "It entails locating and educating people with AI competence, modifying workflow to optimize AI utilization, and integrating new AI tools into the present technological stack.

"In order to support their daily work in a safe manner, CIOs must make sure that their workforce is appropriately trained to design, build, and use Gen AI tools, such as copilots."

Upskilling workers whose jobs could have been affected by Gen AI guarantees that they can concentrate on higher-value tasks and opens up new options for them inside the company. It is important for the leadership to support the effort to integrate Gen AI widely throughout the organization.

"CIOs must create and modify procedures to record the critical knowledge held by employees, as well as update processes to incorporate Gen AI tooling to add value," he concludes.

In order to monitor and reduce new kinds of risk, CIOs must also make sure that their risk management and compliance procedures are updated.

"Lastly, from a technology standpoint, developing, constructing, and expanding Gen AI tools will enable the organization to adapt to its changing needs in a secure and managed way."

Leave a Comment
logo-img Fintech Newz

All Rights Reserved © 2024 Fintech Newz