Incident management · Production

Lyzr AI workflow predicts incident risk and generates remediation steps for NTT Data, saving 70,000+ hours

The problem

NTT Data's change management engineers were overwhelmed by thousands of daily change log entries that manual review could not keep pace with, leaving incident risks unpredictable and remediation guidance inconsistent and slow to surface.

Workflow diagram · grounded in source
1
Engineer submits change log
trigger
“Change management engineers submit logs via a ReactJS UI”
2
ML incident risk scoring
ai_action
“A machine learning model analyzes logs, generating a confidence score on incident risk”
3
Historical similarity search
ai_action
“Qdrant powers similarity search across historical incidents, surfacing relevant past cases and resolutions”
4
AI recommendation generation
ai_action
“A self-deployed GPT-4o-mini model transforms predictions and historical data into actionable incident descriptions and remedial steps”
5
Secure access control
integration
“Okta SAML integration ensures secure, role-based access”
6
Remedial steps delivered to engineer
output
“Engineers receive AI-generated next steps, improving response times and minimizing incident fallout”
Reported outcome

The AI workflow saved over 70,000 hours, improved work-life quality for engineers, and enhanced safety and compliance by automating risk assessment and delivering AI-generated remediation steps.

Reported metrics
Hours saved70,000+
Work-life qualityEnhanced
Safety and complianceImproved
Reported stack
ReactJSQdrantGPT-4o-miniMicrosoft Azure Cosmos DBMicrosoft AzureOkta SAMLLyzr Agent Platform
Source
https://www.lyzr.ai/case-studies/global-it-giant/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The AI workflow saved over 70,000 hours, improved work-life quality for engineers, and enhanced safety and compliance by automating risk assessment and delivering AI-generated remediation steps.

What tools did this team use?

ReactJS, Qdrant, GPT-4o-mini, Microsoft Azure Cosmos DB, Microsoft Azure, Okta SAML, Lyzr Agent Platform.

What results were reported?

Hours saved: 70,000+; Work-life quality: Enhanced; Safety and compliance: Improved (source-reported, not independently verified).

How is this incident management AI workflow structured?

Engineer submits change log → ML incident risk scoring → Historical similarity search → AI recommendation generation → Secure access control → Remedial steps delivered to engineer.

Related incident management cases