Relvy AI

What do they actually do

Relvy AI provides a web app that connects to a team’s observability stack (logs, metrics, traces), dashboards, and code repositories, then generates an interactive “debugging notebook” for each incident. The notebook aggregates relevant signals, runs AI analysis to surface likely causes and next steps, and lets engineers run checks, add findings, and share the investigation. It doubles as an executable runbook that can be reused after resolution (docs – intro, data sources, platform).

Relvy can be used as a hosted service or self‑hosted on a customer’s infrastructure (Kubernetes/Helm), with guidance on architecture and requirements. It needs read‑access to observability data and code to provide useful results, and setup is required to wire in the right sources (self‑hosting docs, data sources). The company is early‑stage (YC Fall 2024; small team), so adoption today likely centers on pilots with on‑call engineers and SREs rather than broad enterprise rollouts (YC listing).

Who are their target customer(s)

• On‑call engineers / SREs: When an alert fires, they lose time hopping between logs, dashboards, traces, and recent code changes to form a hypothesis. They want a single place that collects context and suggests concrete next checks so they can act faster (intro).
• Small reliability or ops teams at startups: With few people covering incidents, repetitive checks and tribal knowledge slow them down. They need a way to capture investigations as reusable, editable runbooks so fixes don’t live only in one person’s head (platform – executable runbooks).
• Engineering managers / incident commanders: Post‑incident notes are often scattered or inconsistent. They need structured, shareable investigations with recorded checks, findings, and actions to support clear handoffs and repeatable postmortems (platform).
• Platform or DevOps teams that own monitoring: They’re tired of building and maintaining custom scripts/connectors for every tool. They want integrations that read logs, metrics, traces, and dashboards without forcing a tool replacement (docs – data sources).
• Security/compliance teams at larger companies: They can’t send sensitive logs or code to third‑party clouds and require SSO, audit controls, and on‑prem options. They look for self‑hosting and enterprise features to meet those constraints (self‑hosting docs).

How would they acquire their first 10, 50, and 100 customers

• First 10: Run targeted pilots with YC startups and existing SRE contacts, offering a 30‑day free trial, white‑glove connector setup, and 1–2 co‑debug sessions to prove value quickly (pricing, data sources, YC listing).
• First 50: Leverage early pilot wins into brief case studies and referral incentives; do focused outreach to platform/DevOps teams at fast‑growing startups and sponsor SRE meetups/Slack communities, while shipping turnkey connectors and templates to reduce onboarding friction (platform, pricing, data sources).
• First 100: Package an enterprise tier (SSO/SCIM, self‑host, audit controls) with repeatable onboarding playbooks; list on observability marketplaces, build partner integrations, and use reference customers to close mid‑market annual contracts (self‑hosting, pricing, platform).

What is the rough total addressable market

Top-down context:

Relvy sits between observability and incident response. Observability platforms are estimated in the low single‑digit billions in 2024, while broader incident response/management (including services) is cited in the tens of billions globally (FMI; Mordor, Mordor, Grand View Research).

Bottom-up calculation:

Using Relvy’s public pricing, e.g., $19/user/month (~$228/user/yr), 100k teams with 20 seats each implies ~$456M ARR; a per‑incident model (e.g., $6/incident) with 100k teams at 200 incidents/yr yields ~$120M ARR (pricing, product post).

Assumptions:

• Average 20 paid seats per team; mix varies by company size.
• Sustained list pricing or similar effective ARPU; discounts may reduce realized ARR.
• Sufficient number of engineering teams operate on‑call workflows to reach 10k–100k paying teams over time.

Who are some of their notable competitors

• Datadog: Full observability + incident tooling with collaborative Notebooks and AIOps (Watchdog/Bits AI); teams already on Datadog can centralize context and AI‑assisted investigations in one place (Notebooks, Event Management/AIOps).
• PagerDuty: Incident orchestration and alerting leader; Event Intelligence and generative AI features handle automated grouping, probable‑cause hints, and diagnostics focused on triage and routing (Event Intelligence, Generative AI).
• FireHydrant: Incident management built around runbooks, service catalogs, and automation; strong when teams prioritize codified playbooks and execution hooks (Runbooks, Docs).
• Blameless: SRE/incident tooling emphasizing runbook automation, postmortems, and reliability workflows—an alternative for teams seeking an opinionated SRE process (Changelog, ServiceNow store).
• Honeycomb: Observability optimized for high‑fidelity queries and fast exploratory debugging; strong for deep query/trace exploration versus Relvy’s AI‑guided, shareable investigation notebooks (Use cases – incident response).