ProductionEvidence: Medium65/100

Saving Sellers over 32,000 Hours of Manual Work Using Amazon Bedrock with US Foods

Use case typeSales enablementUpdated Jul 15, 2026

US Foods built the Automated Order Guide generative AI application on Amazon Bedrock to reduce manual work and help sales teams personalize sales pitches. The solution uses Amazon Bedrock with Anthropic Claude and Amazon Textract, was piloted before production, and was rolled out to 3,300 sellers.

Organization: US Foods
Industry: Consumer & Food
Location: United States
Published: July 2026

Reported outcomes

10x

return on investmentOther quantified impact

32,000 hoursmanual work saved3-4%proposal response time6 weekstime to proof of concept

Strategic outcomes

Other strategic outcomeGave sellers more time for customer-facing workCost efficiencyPlanned embedding into the sales system

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Manual work saved: 32,000 hours decrease

AWS Solutions Case StudiesJul 15, 2026Customer storyExplicit claimMedium evidence strength

saved over 32,000 hours of manual work within 6 months

Normalized claim

Proposal response time: 3-4%

AWS Solutions Case StudiesJul 15, 2026Customer storyInferred claimMedium evidence strength

Response times per proposal fell from 3–4 hours to 20 minutes

Normalized claim

Time to proof of concept: 6 weeks

AWS Solutions Case StudiesJul 15, 2026Customer storyExplicit claimMedium evidence strength

In 6 weeks, the company completed a proof of concept using Amazon Bedrock

Normalized claim

Return on investment: 10 x increase

AWS Solutions Case StudiesJul 15, 2026Customer storyExplicit claimMedium evidence strength

US Foods has increased its return on investment by 10 times since deploying the application

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: US Foods
Provider: AWS
Maturity: Production
Linked source: AWS Solutions Case Studies

The solution uses Amazon Bedrock with Anthropic Claude and Amazon Textract, was piloted before production, and was rolled out to 3,300 sellers

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1Sales enablement
2Document processing automation
3Content generation

Sales reps spent up to 4 hours per proposal manually reviewing diverse data sources including paper menus, spreadsheets, and handwritten notes in multiple languages.

Built the Automated Order Guide generative AI application on Amazon Bedrock, using Amazon Textract to extract text, handwriting, layout elements, and data from scanned documents, then expanded from a pilot into production for sellers.

Technologies

Amazon Bedrock Amazon Textract Anthropic Claude 3

Response times per proposal fell from 3–4 hours to 20 minutes; sellers saved over 32,000 hours of manual work within 6 months; US Foods reported 10x ROI and planned to embed the tool into its sales system.

Architecture

US Foods built the Automated Order Guide on Amazon Bedrock for generative AI recommendations, using Anthropic Claude in Amazon Bedrock for relevant food and service recommendations and Amazon Textract to extract text, handwriting, layout elements, and data from scanned documents. The application moved from a 6-week proof of concept to a pilot and then into production for 3,300 sellers.

Sources & evidence1

Evidence: Medium65/100Evidence strength