Normalized claim
Team hours saved annually: 40,000 hours/year decrease
projected 40,000 team hours saved annually
Rocket Mortgage: Generative AI contact-center analytics using Amazon Transcribe Call Analytics, Comprehend, and Bedrock
Use case typeCustomer experience analyticsUpdated Jun 13, 2026
Rocket Mortgage, America’s largest retail mortgage lender, built Rocket Logic – Synopsis on AWS to improve client interactions and operational efficiency in servicing contact-center operations. The solution automates post-call transcription and analytics, generates concise call summaries and actionable insights, and uses sentiment and entity extraction to support call resolution and customer self-service.
- Organization
- Rocket Mortgage
- Industry
- Finance
- Location
- United States
- Published
- September 2024
Reported outcomes
First-call resolution increase: +10%
Customer experience
First-call resolution hours saved annually: Approximately 20,000 hours/yearServicing calls deployed: 30,000 calls
Catalog median for customer experience deployments: +25% across 52 reported metrics. Compare benchmarks →
Planned next steps
- The source says the organization aims to achieve Team hours saved annually: 40,000 hours/year.
- Projected savings of nearly 40,000 team hours annually from automating call transcription and sentiment analysis.
Why do we believe this?Outcome claims, sources, and evidence checks
Normalized claim
First-call resolution hours saved annually: 20,000 hours/year decrease
saving approximately 20,000 team member hours annually
Normalized claim
First-call resolution increase: 10% increase
there has been a nearly 10% increase in first-call resolutions
Normalized claim
Self-service share: 70% increase
Approximately 70% of servicing clients fully self-serve over Gen AI powered mediums such as IVR
Normalized claim
Servicing calls deployed: 30,000 calls
Rocket started with 30,000 servicing calls in 10 days
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- Rocket Mortgage
- Provider
- AWS
- Maturity
- Production
- Linked source
- AWS Machine Learning Blog
Rocket Mortgage, America’s largest retail mortgage lender, built Rocket Logic – Synopsis on AWS to improve client interactions and operational efficiency in servicing contact-center operations
Customer identity supportedSource describes one deploymentMaturity supported
Primary read
Use case focus
Showing 3 of 4
- 1Contact Center Analytics
- 2Customer Service
- 3Speech Analytics
- Rocket Mortgage partnered with AWS to deploy AWS Contact Center Intelligence, branded internally as Rocket Logic – Synopsis.
- The workflow uses Amazon Transcribe Call Analytics to transcribe calls, Amazon Comprehend for sentiment analysis and entity extraction, and Amazon Bedrock with Anthropic Claude models to generate call summaries and actionable insights.
- The architecture uses AWS Step Functions with Amazon S3-triggered automation and applies PII redaction, encryption, AWS KMS, and IAM access controls for secure processing.
- The team fine-tuned Anthropic Claude 3 Haiku on Amazon Bedrock for call classification and data extraction.
- Approximately 10% increase in first-call resolutions, saving about 20,000 team member hours annually.
- About 70% of servicing clients fully self-serve through GenAI-powered mediums such as IVR.
Architecture
A fully automated post-call analytics pipeline ingests audio files into Amazon S3, triggers AWS Step Functions, transcribes calls with Amazon Transcribe Call Analytics, stores transcripts for downstream BI processing, redacts PII, applies encryption and access controls with AWS KMS and IAM, and uses Amazon Comprehend plus Amazon Bedrock (Anthropic Claude) to extract sentiment, entities, summaries, and actionable insights.
Sources & evidence1
Evidence: Medium50/100Evidence strength
- Customer explicitly identified
- Deployment status explicitly supported
- Quantified outcome available
- Technical implementation details available
ExpandedExpanded
The same organization appears in newer AI deployment evidence.
- Same organization re-documented as recently as 2026.
Measures whether this deployment's public evidence persists — not whether the system is still in production.
Type: Blog PostPublished: Sep 23, 2024Publisher: AWSEvidence: VendorConfidence: Medium
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Customer evidence
Provider evidence
AI patterns
Browse the catalog
Was this useful?
Community
Comments
No published comments yet.