By Ariel Junqueira-DeGarcia, Managing Director, Broadridge
Since coming to the scene in November 2022, generative AI (GenAI) has been revolutionizing how we work and pushing companies to explore its benefits. But GenAI, which is capable of producing new content — like images, text, or music — isn’t without its limitations and risks.
Tools like ChatGPT might seem like they understand user commands, but they don’t actually understand the content. GenAI merely predicts what’s likely to come next based on patterns. Critics warn against blindly trusting GenAI, which can “hallucinate” by making things up. Content that comes from GenAI needs to be checked by a human. Additionally, security is another concern, as many companies fear inadvertently leaking sensitive information to these systems and, in turn, have restricted or prohibited their use in day-to-day operations.
Despite the shortcomings, the momentum behind GenAI adoption is strong. In Broadridge’s 2024 Digital Transformation and Next-Gen Tech Study, 45% of firms allow staff to use GenAI tools for work purposes, and another quarter are training staff on how to use them. To stay competitive, companies need to find use cases that use GenAI to its fullest potential while also mitigating risks.
Finding opportunities
Identifying practical GenAI use cases isn’t always straightforward. Legacy systems, data quality issues, and legal constraints all pose challenges. Even small-scale AI projects can consume resources with little payoff. A strategic approach is essential.
Companies can start by targeting low-hanging fruit by looking for areas where GenAI can augment existing workflows or automate processes. Take client onboarding: firms currently rely on Know Your Customer ops teams to spot high-risk clients by manually sifting through red flags. AI can do this more accurately and quickly.
Another area is trade capture, where firms sometimes struggle to incorporate news headlines into algorithms, relying on standardized data inputs to convert text information into scores. GenAI models trained on large datasets of news and market data can quickly and accurately incorporate headlines into algorithms, leading to more informed decision-making.
Generative AI can also help firms deal with regulatory changes and upped reporting requirements. With a tsunami of regulatory changes such as T+1, CSDR, and EMIR Refit Phase 2 on the horizon, firms have plenty of opportunities to capitalize on this technology. One example is the move to shorten the U.S. securities settlement cycle from T+2 to T+1. The change requires significant re-engineering of existing exception management practices. Deep learning models can help streamline the settlement process. There are also potential use cases for AI in other parts of the trade life cycle, such as collateral management, settlements, and regulatory reporting.
A three-step plan to get started
So, which project is the best to get started with? Before diving into GenAI projects, companies need to ensure that any implementation aligns with long-term strategies and business goals. To help narrow down the field, there are three points that successful projects need to fulfill.
What: Identify the business case for AI adoption. Identify manual tasks that can be automated or pain points that can alleviated with AI. Out of those projects, identify which can be implemented quickly, in weeks rather than months.
Why: Describe the measurable benefits of implementing AI by identifying quantifiable metrics and measurable business outcomes to measure success.
How: Evaluate the cost and feasibility of the project. Gather information on what data is needed to train the model, how the prototype would be implemented and tested, and whether to use a third-party vendor or build in-house.
By embracing generative AI and adopting a strategic, risk-mitigating approach, companies can unlock a competitive edge, streamline operations, navigate regulatory landscapes, and ultimately, shape the future of their industry. This is not just about staying afloat; it’s about harnessing the power of AI to thrive in the years to come.
