About Sherlocks AI
An AI-powered SRE platform that investigates production incidents autonomously, so your engineers can focus on building instead of firefighting.
2025
Founded
$0.9M Seed
Funding
4.9/5
G2 Rating
28
Reviews
7+
Active Pilots
SOC 2 Type 2
Compliance
Founded by Gaurav Toshniwal (CEO) and Akshat Jain (CTO), both ex-CTOs who spent years dealing with on-call rotations and incident management before building the platform they wished they had.
What Sherlocks AI Does
When an alert fires, Sherlocks dispatches specialized AI agents that investigate the issue, correlate signals across your stack, and deliver a root cause analysis with clear next steps.
70%
Reduction in downtime
90%
Reduction in alert noise
~22 min
Average MTTR (down from ~3.5 hours)
16+
Specialized AI agents
How It Works
Context
Sherlocks learns your system topology, service dependencies, and incident history.
Alerts
Receives alerts from your existing monitoring and alerting tools.
Planning
Agents plan the investigation strategy based on the alert type and system context.
Investigation
Agents query Watson (deployed inside your VPC) to gather data from your infrastructure securely.
Resolution
Delivers a root cause analysis with recommended next steps and a full investigation trail.
The Sherlocks Agent Squad
Each Sherlock is a specialized AI agent trained for a specific domain. They work together to investigate incidents from every angle.
Database
Query performance and schema analysis
Kubernetes
Pod health and resource monitoring
Incident Commander
Response coordination and timeline
Code Analysis
PR and commit root cause analysis
Infrastructure
Cloud metrics and saturation checks
Log Analyst
Pattern matching and anomaly detection
Network
Traffic analysis and connectivity
Security
Vulnerability and access scanning
Cost
Resource optimization and billing
CI/CD
Pipeline failures and deployment health
IAM
Permission and role auditing
Compliance
Policy drifts and audit trails
Performance
Latency and throughput analysis
API Gateway
Endpoint health and error rates
Chaos
Resilience testing and fault injection
On-call
Schedule management and routing
Integrations
Sherlocks connects to your existing stack with read-only access. No agents to install, no infrastructure changes needed.
Cloud Providers
- AWS
- Google Cloud
- Microsoft Azure
Kubernetes
- Kubernetes
- Helm
Observability
- Prometheus
- Datadog
- New Relic
- Sentry
- Coralogix
Databases
- MySQL
- PostgreSQL
- MongoDB
- Redis
CI/CD
- GitHub Actions
- Jenkins
- Azure Pipelines
Logging
- Elasticsearch
- Coralogix
- Loki
Messaging
- Kafka
- RabbitMQ
- SQS
- Azure Service Bus
Collaboration
- Slack
- Microsoft Teams
Security and Data Privacy
Your data never leaves your network. Watson runs inside your VPC with strictly read-only permissions.
Read-Only Access
All integrations use read-only permissions. Sherlocks cannot modify your infrastructure, databases, or applications.
No Raw Data Access
Sherlocks operates on metadata and metrics only. PII is redacted at the source by Watson before any data is transmitted.
SOC 2 Type 2 Certified
Independent third-party verification of our security controls. Trust portal available at sherlocksai.trust.site.
Encryption Everywhere
TLS 1.3 for all communications in transit. AES-256 encryption at rest with separate keys per customer.
Air-Gapped Deployments
For regulated industries, the entire Sherlocks stack can run within cloud accounts owned by you. No external dependencies.
In-VPC LLM Option
AI processing can run in your own cloud tenant using state-of-the-art models. No data sent to external LLM providers.