Incident Root Cause Analysis

Generative and agentic tools analyze logs and incident data to help teams find the root cause of service outages and security incidents.

Business impact

• Time to resolution — DigitalOcean reported a 39% improvement.
• Time to complete critical-system incident analysis — NTT DATA reported 30 minutes instead of three days for five engineers.

Data requirements

• Service logs (text) — Used by DigitalOcean to analyze operational incidents.
• Server logs (text) — Used by Hugging Face during breach response.
• Incident records (text) — Implied by the critical-system analysis described for NTT DATA.
• Internal operational data (text) — Used for Nokia’s internal root-cause analysis work.

AI methods and techniques

• Log analysis — The AI reviews service or server logs to identify likely root causes.
• Agentic workflows — DigitalOcean described GenAI agents being used during operational incidents.
• Generative language analysis — NTT DATA and Nokia describe generative AI use for incident investigation and root-cause analysis.
• Anomaly detection — Hugging Face’s own anomaly detection identified the intrusion that triggered the incident response.

AI models and model families

Codex, local large language model

Industries

Real-world evidence

4 documented case studies on record.

Companies using this: DigitalOcean, Hugging Face, NTT DATA, Nokia.