What challenges did Lumeris face before AI implementation?

The People and Culture team spent significant time answering employee HR questions, creating overhead for HR staff. Employees often waited 1–2 days for responses to policy questions.

Source

Proof of conceptEvidence: Medium65/100

Lumeris Elevates HR Operations with a Generative AI Chatbot Using Amazon Bedrock

Use case typeKnowledge managementUpdated Jun 13, 2026

Lumeris built a generative AI chatbot called Ask P&C to answer employee HR questions about benefits, time off, paid holidays, insurance plans, and other HR policies. The chatbot uses a RAG architecture with Amazon Kendra indexing P&C documents for retrieval and Amazon Bedrock with Claude 3.5 Sonnet for response generation. Lumeris began a proof of concept in January 2024, tested it with the P&C team, and rolled it out to employees in March 2024.

Organization: Lumeris
Industry: Healthcare
Location: United States
Published: May 2026

Reported outcomes

1-2 days

quantified impactTime & speed

90%accuracy

Strategic outcomes

Customer experience & trustImmediate employee HR supportScale & capacityReduced HR inquiry workloadNew product / capabilityDeployed an HR support chatbotSpeed & agilityAccelerated AI solution rollout

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

Normalized claim

Quantified impact: 1-2 days

AWS Solutions Case StudyMay 27, 2026Customer storyInferred claimMedium evidence strength

Employees receive immediate responses for most questions instead of waiting 1–2 days.

Normalized claim

Accuracy: 90%

AWS Solutions Case StudyMay 27, 2026Customer storyInferred claimMedium evidence strength

The chatbot achieved an accuracy rating of over 90%.

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

Customer: Lumeris
Provider: AWS
Maturity: PoC
Linked source: AWS Solutions Case Study

Lumeris began a proof of concept in January 2024, tested it with the P&C team, and rolled it out to employees in March 2024

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1HR assistant
2Knowledge management
3Employee self-service

The People and Culture team spent significant time answering employee HR questions, creating overhead for HR staff.
Employees often waited 1–2 days for responses to policy questions.

Lumeris built 'Ask P&C' as a retrieval-augmented generation chatbot.
Amazon Kendra indexes multiple HR document formats and uses thesaurus mapping such as P&C to People and Culture to improve retrieval.
The solution sends the user query plus retrieved context to Claude 3.5 Sonnet in Amazon Bedrock to generate answers.

Technologies

Amazon Bedrock Amazon Kendra Claude 4.5 Sonnet

The chatbot processes dozens of weekly inquiries that previously required HR staff effort.
Employees receive immediate responses for most questions instead of waiting 1–2 days.
The chatbot achieved an accuracy rating of over 90%.
Lumeris moved from ideation to rollout in 3 months.

Architecture

A RAG chatbot architecture retrieves HR policy context from an Amazon Kendra knowledge base, then passes the query and retrieved context to Claude 3.5 Sonnet in Amazon Bedrock to generate responses. The solution includes thesaurus-based term mapping and supports multiple HR document formats.

Sources & evidence1

Evidence: Medium65/100Evidence strength