ZeroOps meets Generative AI!

Keri, Clara and I have experienced AI pushing the boundaries of ZeroOps over the last several years across multiple use cases that we have helped to deploy at leading global banks. We are now at a true inflection point – where Generative AI has unleashed greater power, lowering the bar to execute on what has so far been custom and complex work. The opportunities for disruption that Generative AI is creating are significant, several are outlined here in Accenture’s recent research report.

With rebalancing human + machine workforce being a core principle of our ZeroOps approach, Accenture has been leveraging AI in financial services operations for several years. In partnership with our clients, we have implemented AI/ML-driven solutions across a variety of use cases and are excited to see that AI is hitting the mainstream. But what is making this moment such an important inflection point? The large language models (LLMs) and foundation models powering generative AI are a big step up from traditional AI and machine learning more broadly. Having solutioned for language complexity, machines can now learn context, infer intent and be independently creative. Unlike discriminative AI that performs a very specific task, generative AI now creates original, varied, and non-repetitive content from training data in various forms including text or code (e.g., ChatGPT), audio (e.g., Jukebox), and visual media (e.g., DALL-E).

While many organizations have started experimenting with “off-the-shelf” foundation models, we are seeing great value of generative AI in the customization of pre-trained models for specific use cases. Generative models can require little data or fine-tuning to be customized to a specific domain or company, enabling organizations to adapt them to their unique needs. Accenture has been helping our clients in FS and beyond to navigate the generative AI landscape with accelerators, real-world insights and best practices on how to effectively design and scale. As we look at the world of Operations, we see broad applicability of Generative AI. Some examples are below:

Use Case 1: Conversational Know-Your-Customer (KYC)

Collecting and confirming a wide variety of data points for KYC purposes can prolong customer onboarding and lead to many frustrations along the way - for clients and employees alike. Generative AI can be leveraged to extract information from structured and unstructured data to build the KYC knowledge required for your organization. With conversational KYC, the human agent can request information easily through a simple chat tool. The agent can leverage ‘embeddings’, upload documents via the chat interface or connect third-party or public data to provide additional information and context to the AI. The AI assistant can generate responses, list the sources for the information used in the response, and even provide a confidence rating for the correctness of their response based on all available data.

Use Case 2: Client servicing - Complaints Operations

Lengthy case times plague complainants and drive costs for Financial Institutions. Generative AI can meaningfully augment complaints operations along the entire lifecycle. For example, by transcribing calls, summarizing previous customer interactions and case history, flagging customer vulnerability indicators, and indicating the sentiment of a client.

Use Case 3: Regulatory Compliance

Staying compliant with frequently changing regulations and balancing varied regulatory requirements across geographies and business lines is a sizable and ongoing challenge for Financial Institutions. Generative AI can be leveraged to identify and close compliance gaps. ?Provided with the text of a new or existing regulation, the AI can review it against your existing documentation on current processes, controls, and procedures and identify gaps or point you to the controls that already address this regulatory requirement. For any identified gaps, it can propose solutions, and even outline a project plan to implement the needed processes and controls.

Use Case 4: Securities Settlement

With its unique ability to digest many data points in the blink of an eye, generative AI is a great assistant in securities settlement processes. Accenture’s ATOM (Applied Technology and Operations for Markets) has been leveraging AI to analyze trade data and identify potential settlement issues, such as incomplete or inaccurate trade data, mismatched trade details, or failed trades. Generative AI is now used to predict settlement failure risks and proactively prevent failures by crafting an e-mail to both parties alerting them to the risk and outlining steps to resolve the issue.

The above use cases demonstrate how generative AI can be leveraged to reduce costs, minimize risks, delight customers, and equip human agents with an AI co-pilot that allows them to focus on the most value-added tasks. Every role in the enterprise has the potential to be reinvented as AI dramatically amplifies what people can achieve. So why have several Financial Institutions recently restricted employee use of ChatGPT? These decisions appear to be driven by compliance concerns and the need to further vet the use of third-party software, not a sign of disinterest in the technology behind these tools. Quite the opposite: Accenture’s CXO Pulse Survey found that over half of organizations are planning pilot cases in 20231. We understand the caution around the usage of publicly available tools and the need for companies to carefully evaluate the legal, ethical, and reputational implications brought on by the adoption of this novel and quickly evolving technology. We encourage organizations to experiment with generative AI, establish Responsible AI principles, and invest in their AI maturity.

For Operations, we believe there are significant opportunities for Financial Institution to embrace Generative AI. The use cases outlined above are only the tip of the iceberg. From customer onboarding to reconciliations, the opportunities to augment core operations capabilities, unlock productivity, and increase customer satisfaction by leveraging generative AI are plentiful.

Now is the time for Financial Institutions to redefine their operations and grow their AI capabilities. If you have any questions or want to discuss how you can leverage the power of generative AI in Operations, please reach out – we have insights to share.

Rajat Dev | Keri Smith | Clara Wienhold