At KPMG, where AI is an essential part of digital transformation, Marisa identifies opportunities for integrating the technology into various aspects of the company’s operations. Her main area of focus is document investigation, wherein specialists are required to review many disparate financial and legal documents to determine risk across a portfolio. Under Marisa’s leadership, KPMG is using IBM Watson Discovery as the backbone to a question-answering pipeline that is built into a specialist workflow by auditors and risk analysts. The platform is delivering results, while also enabling KPMG to harness data that’s critical to its modernization journey.

What was the business opportunity you sought to address by using AI?
Our organization is built upon professionals performing analyses under heavy regulatory oversight, which means that any digital transformation must be introduced carefully. Our end goal must be to augment our professionals to do the strategic work that resists automation, and that necessitates the use of AI. In the meantime, there are a host of technologies we can make use of to take steps toward modernization, and many times we can integrate AI at critical steps to help further the process along. Finding those niches is the main challenge in my work.

Please tell us more about your project and how you’re using Watson in your business.
My main area of focus is document investigation. I find this area fascinating, not only because of my technical background in question-answering (and applying Watson to healthcare), but also because the methods of investigation involve iterative question-answering by a variety of experts to reach a final conclusion. This requires a rethinking of typical methods of using technologies. Because of the nature of the things being investigated (various text documents), the nature of the investigation itself (iterative question-answering), and the nature of the people who do the investigation (specialists with years of experience), Watson Discovery enables us to deliver results while collecting the critical data needed to understand modernization.

What makes an AI project successful?
First, understand the full challenge trajectory. You are rarely the first person to have tried to solve it. What have others done beforehand? Why did they fail? What succeeded? Second, lead a united front. If this challenge is truly worth AI, then it will need subject matter professionals who are an integral part of the solution, not just people you call on with specific questions. Similarly, if this challenge can truly be solved right now, it will need to live in an ecosystem of people and processes that already exist. Those people and processes must be part of your project. Third, know that you’re not going to solve the challenge today. One of the things I find most rewarding about the work I do is taking the time with my team to figure out the most elegant and beautiful solution, even if it’s just to know it’s out there. The steps required to get there might be a slog, and there might need to be some inelegant stepping stones built in the meantime. But keeping hold of that vision of the solution to the larger challenge, while solving today’s not-so-interesting challenge, is a key way to find success in a messy world.

What advice would you give your younger self? 
I wish I had better understood that the “correct” technical solution provided at the wrong time is not always the right solution. I spent many years in research learning to craft those correct solutions, but not understanding why they never seemed to make it out in the real world. They were correct, the absolute best way to do something, but just couldn’t survive yet in the existing business environment. On the other hand, thinking through the correct way to do something is never a lost cause. It not only allows you to plan for the future state, but to better understand the surrounding environment and circumstances needed to make it happen, and start chipping away at changing those structures. I learned many years ago from one of my favorite mentors that the trick to being a career technologist is learning when to shelve your pet project, knowing that eventually you can dust it off and pick it up again when the time is right.

This article represents the views of the author(s) only, and does not necessarily represent the views or professional advice of KPMG LLP.

The information contained herein is of a general nature and is not intended to address the circumstances of any particular individual or entity. Although we endeavor to provide accurate and timely information, there can be no guarantee that such information is accurate as of the date it is received or that it will continue to be accurate in the future. No one should act upon such information without appropriate professional advice after a thorough examination of the particular situation.

She puts Watson Discovery to work

Dive deeper into your journey to AI and learn more about Watson Discovery