US Foods serves about 250,000 customer locations and posted $39.4 billion in fiscal 2025 sales. Its growth depends on three things: how many accounts sellers can call on, how much each account buys, and how cheaply orders get delivered. The company now has an AI tool aimed at each, and the disclosures around each one differ.
US Foods’ AI Proposal Tool Saves Sellers 32,000 Hours
The tool, called the Automated Order Guide, reads a prospect’s menus and notes and builds the product proposal, and it saved more than 32,000 seller hours in its first six months, the AWS case study said. US Foods engineers said at the AWS re:Invent conference in December 2024 that onboarding work that took three to four hours now takes 20 to 30 minutes.
Dave Flitman, chair and CEO of US Foods, said on the second-quarter call that when AI handles the paperwork, sellers get time to find the next customer and sell more lines to the ones they have. Lines per customer have risen for several quarters, he said, and he credited part of that to AI support for sellers.
US Foods has not isolated how much of that improvement came from the tool. Seller headcount rose 8% in the quarter, so the claim is more output per seller rather than fewer sellers.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
Visit Assistant Delivers 700,000 Sales Prompts in Six Weeks
Visit Assistant Insights, which US Foods built in-house, reads customer-specific data and produced more than 700,000 prompts on which accounts to prioritize and what to raise on the visit in its first six weeks of use, Flitman said on the Aug. 6 call.
Existing customers order repeatedly, so a larger share of each order is recurring revenue with no acquisition cost. US Foods has not said how many of the 700,000 prompts led to purchases or how much revenue they produced. That is the figure finance chiefs now want. According to the PYMNTS Intelligence report “What Happens When CFOs Get Serious About Gen AI,” based on a December 2025 survey of 60 CFOs at U.S. companies with at least $1 billion in revenue, 56.7% judged generative AI returns by revenue per customer, up from 18.3% five months earlier.
Descartes Routing Lifts US Foods’ Cases per Mile 2%
Descartes routing software, fully deployed across US Foods’ network in 2025, improved cases delivered per mile in broadline delivery by about 2% from 2024, the company showed in its fourth-quarter presentation.
AI-driven demand forecasting and labor planning are improving service and productivity while reducing working capital, Flitman said on the August call. The routing figure is one of two AI results the company has tied to a stated operating metric; the other is a 3% lift in cart conversion from AI search in its MOXe ordering platform, disclosed in November 2025. The rest it folds into a 3% to 5% annual productivity target without breaking AI out.
Large companies elsewhere report the same mix of positive early returns and a long road to full payback. According to the PYMNTS Intelligence report “The Enterprise AI Payback Curve: Adoption Accelerates as Returns Take Shape,” based on a June survey of 60 senior technology executives at U.S. enterprises with at least $1 billion in revenue, 87% said generative and agentic AI tools performed better than expected on competitive position, between 90% and 100% rated their recent return on investment as positive, and 50% to 63% put full payback at five to six years. None expected to cut spending.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
Add as Preferred Source
