Citi transforms critical internal audit with machine learning, NLP and AI
IBM Watson® Discovery, IBM Cloud Pak® for Data and IBM® OpenPages® with Watson empower auditors to better serve customers
At first glance, the world of documentation reviews and risk assessments wouldn’t appear to be the next big hot spot to innovate with the newest and shiniest data and AI tools. But at Citi, Marc Sabino is building a practice he calls audit of the future, where cutting edge machine learning, natural language processing (NLP) and advanced analytics solutions available through IBM Watson Discovery, IBM Cloud Pak for Data and IBM OpenPages with Watson will forever transform the daily work of the bank’s 2500 auditors.
Internal audit: An innovator’s playground
Being coined head of innovation for internal audit says it all. Sabino embarked on his mission almost four years ago after a long tenure at Citi spanning well over two decades. Ironically, none of those roles involved auditing, but internal audit housed a surprising wealth of data from across the company. To Sabino, the practice was just the place to use data proactively with more predictive capabilities than ever before seen elsewhere at Citi.
Sabino and his team started to re-imagine the role AI and natural language processing could play in an auditor’s workflow. They thought about how they could use that technology to scale control testing by applying it to much larger data sets than before to uncover themes and insights.
Sabino was also thinking through how to bring machine learning and AI into the full audit lifecycle to find anomalies within business monitoring, then plan and scope audits more effectively using all the fresh findings. He geared up to give auditors superpowers with greater detection capabilities and beefed-up data-driven insights—all within a unified platform that could enhance stakeholder engagement and collaboration.
In his own words
—Marc Sabino, Head of Innovation of Internal Audit, Citi
Not just nice to have: Building AI into workflows
Citi has one of the largest corporate audit departments in the world. Given the scale and importance of the function, any innovation Sabino introduced would have a high degree of scrutiny because the auditors are also responsible for reporting their activities and outputs to the regulators. Replacing the bank’s standing audit platform would require careful consideration and a high degree of trust in a technology partner. Having become familiar with Watson tools, Sabino approached IBM with three requirements: He wanted a new audit platform injected with advanced analytics and AI; it must be easy to use; and it needed to be the best in the industry.
Learn how to automate and accelerates business processes based on insights from financial documents. Visit: https://www.ibm.com/cloud/ai/ai-document-insights
How IBM accelerated Citi’s AI journey
In January 2020, IBM led a design thinking workshop that brought together five IBMers and 25 Citibankers. That was enough to both engender trust and set forth a vision. From there, a project emerged to build a solution involving people, processes and platform that could save hundreds of thousands of billable hours. The vision was differentiated enough yet realistic to achieve that Citi selected to move forward with the solution and deployment.
A significant force of IBM technical talent and subject matter expertise was brought onboard to support the project:
- IBM Data Science and AI Elite Team helped identify use cases and demonstrate IBM Cloud Pak for Data’s value
- IBM Expert Labs skillfully eased the integration with Citi’s IT organization
- Promontory, an IBM group, tackled regulatory requirements.
Eliminating manual processes and improving data quality
Initially, three projects were spawned that harnessed a broad array of IBM Data and AI solutions available through IBM Cloud Pak for Data and IBM Open Pages with Watson. The first use case gave auditors access to tooling that analyzes the control definitions and scored them based upon the 5 “W”s (Who, what, when, how and why). A second project was designed to help auditors tap into Watson Discovery to save time processing thousands of transcripts between agents and customers monthly. A third use case is now underway to bring IBM Watson® Assistant to the task of assisting auditors with identifying relevant content from within hundreds of pages of audit manuals.
And the list goes on. That’s because IBM has helped create an AI innovation space with Sabino’s team so they can keep innovating using Watson tooling on their new auditing platform. By landing these various short burst proof of concepts, Citi has gained confidence in IBM, not only as a partner for audit transformation, but also as an innovator in AI.
Watch this demo to see how IBM Open Pages with Watson can facilitate effective controls management.
Address multiple data and AI use cases with a wide array of integrated, cloud-native data and AI services. Visit: ibm.com/products/cloud-pak-for-data/services