Speaker: Eli Gill
Please briefly introduce yourself.
Eli Gill: Hi, I'm Eli Gill. I'm the Vice President of Product Engineering and AI at Paro. Paro is a Chicago based company that matches talent on one end, uh particularly in the finance and accounting space with companies who have the needs for that talent. We use a unique blend of data driven methodology plus AI to help perform that matching. And uh in my career, I've worked with AI for over a decade now and I'm excited to be presenting um in the upcoming Finance Alliance meeting.
What will you be speaking about at the event?
Eli Gill: For the upcoming event. I will be hosting a workshop called uh AI and the CFO separating promise from reality. The nature of the talk is going to be trying to dig in, find where the state of AI is right now. What limitations it has, what challenges, what challenges um can be overcome and how you can deliver value to your business with AI. I think the fun part of this uh is it, it's a workshop. So as part of that, you will have access to a demo environment, you'll be able to go in, interact with a real AI and a real data set. So you can see not only where limitations are but where you'll be able to provide value.
3. What are some of the practical applications for AI in finance you’ll be demonstrating?
Eli Gill: In terms of practical application, we will be diving into a couple test data sets and we will show how if your data is structured properly, you'll be able to ask an AI assistant any question within that data set and get an answer. We'll show not only where you can get relevant and contextual answers for your business, but also where sometimes the AI algorithm is weak and how you can improve that weakness.
What is one thing companies often get wrong about AI for finance?
Eli Gill: Uh in terms of what companies get wrong about AI or like, where are their common misconceptions? Uh I would say the following - AI doesn't have one generally well accepted definition. I think if you ask anyone, um they would give you a slightly different answer than anyone else. However, even within that, um there are some areas where people use the term AI and it's probably a little bit outside of the scope of what most people would say AI is. AI and machine learning are generally ways to um interact with uh algorithms that have been developed to aid in some form or fashion, whether it be getting um predictions or whether it be asking a question and getting an answer where I commonly see people uh conflate the term AI with other things is automation. Um very often as I'm working with um clients, prospects, ect., Um When they think AI, they think um more of like a robotic automation and uh I would generally put that as separate and distinct from AI. I think another common misconception um is AI is gonna replace everyone's job. I don't think that will be true. I think AI is going to become something that people will use as like a copilot or a concierge service to help them be more productive throughout their day.
What are some of the key challenges finance leaders face when implementing AI for the finance function?
Eli Gill: When it comes to implementing AI uh for a finance and accounting team, I think there are really two things that I would hone in on as challenges First, Uh focus on what AI is good at doing versus what it's not good at doing. I'll talk about this during um my workshop. But uh if you're going to go interact with a chat bot, it's probably not going to be great at, let's say, doing some math for you. So if you expect it to help you with any sort of um formula that you might have within your business, the chat bot is really not going to be the best tool for you to use. You'll probably want to build a different type of algorithm that can help you answer those kind of questions. So making sure that you're aligning your expectations with what the algorithm can deliver - very common challenge. I think number two is understanding that um while we want to measure ROI, while we want to measure outcomes and we certainly should. Um not every AI or machine learning project is successful. Um There are times when after the models are trained after everything is done after it goes into production, it doesn't quite give the performance that we thought. So while finance and accounting leaders always want to see a tangible benefit right away, they do need to be prepared to face models, not performing to the standard they thought and um iterating on that model building on top of it and making it what they want to see in terms of ROI.
