Sherlocks.ai
AI-powered SRE platform — 16+ specialized agents investigate incidents and deliver root cause to fixes in minutes, 24/7.
Overview
Sherlocks.ai is an AI-native site reliability engineering (SRE) platform that deploys 16+ purpose-built AI agents — covering domains such as Database, Kubernetes, Log Analyst, Network, CI/CD, Security, and On-call — to autonomously triage production incidents, correlate signals across the full observability stack, and deliver ranked root-cause hypotheses with remediation steps directly inside Slack. Its proprietary Awareness Graph links live telemetry with historical incident records, resolved tickets, runbook changes, and past Slack conversations so each investigation builds on institutional memory rather than starting from scratch. The platform connects to 40+ tools (AWS, GCP, Azure, Datadog, New Relic, Prometheus, PagerDuty, GitHub Actions, ELK, Kafka, and more) with read-only access, installs via a Helm-deployed Watson agent in your own VPC, and supports three deployment models: SaaS, Cloud-Native, and fully Self-Hosted with optional private LLM inference.
The company reports that teams using Sherlocks reduce alert noise by 90% and cut mean time to resolution from roughly 3.5 hours to around 22 minutes, a claimed 70% reduction in downtime. Full autonomous capability builds progressively over the first two to three months as the Awareness Graph ingests incident history, meaning teams should expect a ramp-up period before seeing peak effectiveness. Sherlocks is SOC 2 Type 2 certified and encrypts all data in transit (TLS 1.3) and at rest (AES-256), making it viable for teams with strict security and compliance requirements.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 31 Aug 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- 30 investigations/month
- All 16+ AI agents
- Slack + Microsoft Teams
- Community support
- Unlimited investigations
- SSO/SAML, audit logs, RBAC
- Dedicated Field Engineer + SLA support
- Air-gapped/in-VPC LLM options
Use cases
What you can produce with Sherlocks.ai
- Root-cause analysis report with ranked hypotheses and remediation steps delivered in Slack within minutes of alert firing
- Continuous 24/7 autonomous monitoring across cloud infrastructure, databases, Kubernetes, CI/CD, and message queues
- Awareness Graph that accumulates and correlates team incident history, runbook changes, and resolved tickets over time
- Alert correlation and deduplication reducing actionable alert volume by a claimed 90%
- Three deployment models — SaaS (in-VPC agent), Cloud-Native, Self-Hosted — with SOC 2 Type 2 and TLS 1.3/AES-256 security
- Integration connectors for a documented set of ~19 systems (AWS/GCP/Azure, Kubernetes, Datadog, New Relic, Prometheus, ELK, Kafka, GitHub Actions, Jenkins, Sentry, and major databases), with custom integrations built on request
- Flexible private LLM inference option via Azure OpenAI, AWS Bedrock, or self-hosted models for data-sensitive deployments
ASEAN Perspective
Sherlocks.ai in Southeast Asia
Sherlocks.ai was co-founded by IIT-educated engineers from India and counts several Indian tech companies (TradeIndia, Fynd, Topmate, Lokal, Stable Money, SpeakX) among its earliest customers, giving it meaningful early traction in the broader South Asia market. The integration stack — AWS, GCP, Kubernetes, Datadog, Prometheus — aligns well with what ASEAN engineering teams at growth-stage startups typically run. However, as of mid-2026 Sherlocks has no stated APAC data residency, no Singapore or SEA office, and support is US time-zone-oriented, which are practical gaps for regulated industries in Singapore, Indonesia, or the Philippines that require local data processing or SLA-backed regional support.
Sherlocks.ai splits SRE work into 16-plus specialist AI agents — one for Kubernetes, one for databases, one for CI/CD, and so on — then has them reason together on an incident and post findings to Slack or Microsoft Teams (both are live integrations, not a roadmap item). The Awareness Graph, which links live telemetry to past incidents and old Slack threads, is the more interesting idea: institutional memory that most incident tools throw away. G2 shows a strong early signal, 4.9 out of 5 across roughly 28 reviews. Pricing is public, unlike most of this category — 30 free investigations a month with no credit card, Enterprise beyond that. The company is genuinely small: a pre-seed round around $900K closed in mid-2026, founded by Gaurav Toshniwal and Akshat Jain. Full autonomy takes 2–3 months to build up as the graph ingests history, so this isn't a day-one win, and most named customers so far are India-based (TradeIndia, Fynd, Topmate).
What people say
G2 puts Sherlocks.ai at 4.9 out of 5 across roughly 28 reviews as of early 2026 — a small sample but a consistent one, with reviewers calling out fast Slack-native root-cause delivery and real-time visibility into incidents. That's the headline strength: 16-plus purpose-built agents (Database, Kubernetes, Log Analyst, Network, CI/CD, Security, On-call, and more) investigate in parallel and hand back ranked root-cause hypotheses instead of a wall of correlated alerts.
The company's own numbers claim a 90% cut in alert noise and mean-time-to-resolution dropping from roughly 3.5 hours to 22 minutes — vendor-reported, so treat as directional rather than audited. What backs it up architecturally is the Awareness Graph, which ties current telemetry to historical incidents, closed tickets, and old Slack threads, so an investigation six months in should be sharper than one on day one — the flip side being that full autonomy takes 2–3 months to build up, not something teams get immediately.
Two things worth correcting about how this product tends to get described: pricing is not quote-only. There's a genuine free tier — 30 investigations a month, no credit card — before Enterprise kicks in with SSO/SAML, audit logs, and air-gapped LLM deployment for regulated teams. And Microsoft Teams is a live integration alongside Slack, not a "coming soon" item, per Sherlocks' own Microsoft Marketplace listing.
At roughly $900K raised in a mid-2026 pre-seed round led by SenseAI Ventures and under 10 people, this is a genuinely early-stage company; named customers so far skew Indian (TradeIndia, Fynd, Topmate). ASEAN teams will recognize the integration stack — AWS, Datadog, Kubernetes — but there's no advertised regional data residency for Singapore or Southeast Asia yet.
Summary of public user & expert reviews, compiled by RECATOOLS.
Notable facts
- Gaurav Toshniwal previously scaled Doubtnut — an Indian EdTech platform — to 50 million users before co-founding Sherlocks.ai, giving him firsthand experience of the on-call pain he set out to solve.
- The platform deploys exactly 16 named specialist agents including a 'Chaos Sherlock,' a 'Compliance Sherlock,' and an 'IAM Sherlock' — covering attack surfaces most generic AIOps tools ignore.
- Sherlocks calls its architecture the 'Awareness Graph,' a living graph that correlates current telemetry with historical Slack conversations and resolved ticket threads — so the AI remembers what your team tried last time a similar alert fired.
- Despite being US-headquartered, Sherlocks.ai's earliest published reference customers are almost entirely India-based, making it one of the few SRE AI platforms with genuine South Asian design-partner roots.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including Sherlocks.ai's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Sherlocks.ai unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Sherlocks.ai directly →
Spotted something out of date? Suggest an update →
More in Code & Dev Tools
