Bechtel’s machine learning models outperformed traditional cash forecasts, happily surprising AT Dan Degagne.
Earlier this year, Bechtel assistant treasurer Dan Degagne spearheaded an in-house project to develop a machine learning (ML) model designed to relieve some of the pain of generating accurate cash flow forecasts—a challenge shared by just about every company in the NeuGroup Network.
- Mr. Degagne had doubts about the ultimate outcome. But in a new NeuGroup Insights video you can watch by clicking the play button below, he says he was pleasantly surprised when Bechtel’s proof-of-concept ML model exceeded expectations by producing a forecast that “was more accurate than some of our bottom-up forecasts that take a lot of time and effort.”
- Behind some of his surprise was skepticism about all the attention and media hype surrounding generative AI. “I was somewhat skeptical going in because there is so much talk about machine learning and AI, especially in the finance space,” he says.
- “Everybody’s trying to sell you the latest or the newest implementation of AI without, it seems, really understanding what’s underneath that.”
The video clip is taken from an upcoming episode of the Strategic Finance Lab podcast, available on Apple and Spotify. In it, NeuGroup’s Justin Jones interviews Mr. Degagne and his fintech lead Shayly Nguyen about the ML project, which is still in an early stage, with much more work required before the model is implemented.
- Mr. Degagne recommends companies avoid having unrealistic expectations. “In almost every conversation now where there’s some corporate problem, particularly in finance, inevitably somebody is going to say, ‘Can AI or ML deal with that?’” he says in the full podcast. “I think there’s a big risk there.”
- While ML offers exciting opportunities, it’s crucial to avoid buying into the hype without a deep understanding of the tools. He said Bechtel’s success so far with its model demonstrates how, with the right approach and expertise, advanced technologies can indeed make a significant impact—but only when applied thoughtfully and cautiously.
