"As Intuit Chief AI Officer Ashok Srivastava told me prior to joining Intuit CEO Sasan Goodarzi and Fortune editorial director Andrew Nusca on stage, relatively few leaders are sharing data on the returns they're getting from AI. (Intuit has quantified AI's bottom-line impact on earnings calls and shared data from customers using its new QuickBooks offering, which I wrote about earlier this year.)"
"Corporate adoption will take time-and then take off. While individuals may embrace AI, companies have to contend with legacy systems, cultural challenges, employee resistance and the reality that pilot projects are often hard to scale. Add in concerns around governance, security, data quality and keeping customers happy. But nothing shakes inertia like a fear of extinction, and the widening lead of AI winners is forcing CEOs to take action, bubble or no bubble."
"Predictive AI remains powerful stuff. As Stephen Messer, cofounder of Collective[i], pointed out yesterday: "If you look at all the people who created the stuff that's now hot, none of them are focused on language models." He's right. Stanford scientist Fei-Fei Li has moved on to building AI with spatial intelligence; NYU's Yann LeCun is building systems based on self-supervised learning and world models, or AI that learns to simulate and predict real-world dynamics, which is critical"
CEOs will face pressure to demonstrate measurable ROI from AI, with increased focus on growth impacts rather than solely cost reduction or headcount cuts. Intuit has quantified AI's bottom-line impact on earnings and shared customer data from its QuickBooks offering as an example of measurable benefit. Corporate adoption will be gradual due to legacy systems, culture, employee resistance, governance, security, data quality and scaling challenges, but competitive pressure and the widening lead of AI winners will accelerate deployment. Attention will expand beyond large language models toward predictive AI, spatial intelligence, self-supervised learning and world models for realistic simulations and what-if planning.
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]