"Goldman Sachs chief information officer Marco Argenti thinks monitoring every employees' AI use is the wrong way to measure whether it's actually making people more productive. As companies increasingly push employees to adopt the technology to try to spur productivity, Goldman's Argenti is taking a different approach by measuring how quickly his engineering teams can move from coming up with an idea to actually executing on it."
"Even though the investment bank can see exactly how much each of its employees uses AI, including the 12,000 engineers Argenti oversees, laser-focusing on that detail isn't helpful, he told Business Insider. "It would be like looking at only one player on the field," Argenti said. "Fine, this player is doing more movements, but why am I not scoring more goals? Well, because they need to pass the ball.""
"What works better, he added, is measuring the speed at which a team can develop a feature, which he said you can see because suddenly a productive team's work backlog starts shrinking rapidly. AI tools have helped the bank's employees go from creating PowerPoints for their ideas to creating prototypes that can be adjusted in real time based on feedback, he added."
""There's zero time between idea and prototype. You kind of '3D print' software," Argenti said. Goldman has been on the forefront of implementing AI for years. In 2024, the company launched its GS AI Platform, which incorporated large language models like those from OpenAI and Google, but with a layer of security to protect the company's private data."
Goldman Sachs leadership views individual AI usage tracking as an incomplete measure of productivity. Instead, engineering performance is assessed by how quickly teams move from an idea to a working feature. Faster development shows up as a rapidly shrinking work backlog. AI tools have reduced the time between concept creation and prototype creation, enabling real-time adjustments based on feedback. Employees have shifted from producing static materials like PowerPoints to generating prototypes that can be iterated quickly. Goldman has built an AI ecosystem including a secure GS AI Platform using large language models, an internal ChatGPT version, and a research assistant called Legend that enables natural-language searching across internal files without knowing where information is stored.
#ai-productivity-measurement #software-development-velocity #enterprise-ai-platforms #engineering-workflow #backlog-reduction
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]