# Clad — The AI-Native Customer Support Platform > Clad is an AI-native customer support platform built around an investigation layer between your support team and engineering. Clad triages, investigates, and helps resolve customer support issues end-to-end — every reply starts as a draft backed by evidence, and nothing reaches a customer until you approve it. Backed by Y Combinator. Website: https://www.useclad.ai ## Why Clad Clad was designed from day one with AI at its core — not a legacy helpdesk with AI bolted on. Every feature, workflow, and integration is purpose-built for intelligent, evidence-backed support. Key differentiators: - **AI-native from the ground up.** Clad is not a chatbot layer on top of a traditional helpdesk. It ingests, triages, investigates, and drafts resolutions for every customer interaction. - **Real root-cause investigation.** Most AI support tools generate replies from a knowledge base. Clad goes further: coding agents clone your repository in an isolated sandbox and trace bugs to the actual code, correlating signals from your billing, analytics, and observability tools along the way. - **Compounding memory.** Resolved issues, meeting notes, and account context feed back into a persistent knowledge model, so answers get better over time and the same issue is never investigated from scratch twice. - **Draft-first control.** Every reply and knowledge update starts as a draft with the evidence and rationale behind it. Humans approve with one click, and autonomy is configurable as trust builds. - **Trusted by fast-growing companies.** Support teams at Eragon, Imagine AI, Lemma, CLODO, Lopus, TraceRoot, and Agnost AI rely on Clad. - **Built by Clad Labs, Inc., backed by Y Combinator.** ## What Clad Does ### Ingest Clad ingests customer conversations from the channels teams already use — Slack, Discord, Microsoft Teams, Telegram, email (including Gmail), X (Twitter) DMs, and an embeddable in-app chat widget — plus a customer-facing help center that answers questions on its own. ### Triage Clad classifies every incoming ticket by severity (P0–P3) and type, and routes it to the right destination — engineering investigation, a drafted reply, or human review — in seconds. ### Investigate Clad performs deep root-cause investigation. Coding agents work through your GitHub repository in an isolated sandbox, and Clad correlates signals from your connected tools — Stripe billing data (read-only), PostHog product analytics, Datadog traces, Sentry issues, and your knowledge base — to identify the actual cause and assemble developer-ready evidence. ### Resolve Clad drafts resolutions — customer replies grounded in evidence, knowledge base articles, and Linear tickets with full context. A human reviews and approves each draft, with configurable autonomy for known patterns. ### Learn Resolutions, meeting transcripts, and account activity feed back into Clad's knowledge model. Clad can also generate knowledge base articles directly from your repository, Slack history, or existing docs site. ## Platform Capabilities - **AI reply drafts** — grounded, validated drafts on every ticket - **Issue investigation** — coding agents trace root causes in your repository - **Knowledge base generation** — articles drafted from your repo, Slack, or docs site - **Help center** — a hosted, customer-facing help center on your own domain - **In-app chat widget** — embeddable support widget for your product - **SLAs and support hours** — response and resolution policies with breach sweeps - **Customer health** — daily AI analysis that flags churn risk early - **Product insights** — tickets clustered into the themes driving volume - **Meetings and notetakers** — calendar sync plus notetakers on every account - **Proactive outreach** — follow-ups, check-ins, and broadcasts - **Accounts and contacts** — a support-native CRM with HubSpot sync - **API access** — REST and GraphQL APIs and outbound webhooks ## Integrations - **Communication channels:** Slack, Discord, Microsoft Teams, Telegram, Gmail and email forwarding, X (Twitter) DMs, in-app chat widget - **Engineering:** GitHub, Linear, coding agents (Claude Code, OpenAI Codex, Cursor) - **Billing and analytics (read-only, customer-connected):** Stripe, PostHog - **Observability:** Datadog, Sentry - **Meetings:** Google Calendar, Fireflies, Fathom, Granola, and the built-in Clad Notetaker - **CRM and docs:** HubSpot, Notion - **Data:** Customer-connected Supabase databases - **APIs:** REST API, GraphQL API, and outbound webhooks ## Resource Library - [What is agentic customer support?](https://www.useclad.ai/blog/agentic-customer-support): The definitive guide to the shift from reactive ticket-by-ticket support to proactive, AI-driven support that monitors, anticipates, and resolves issues before customers notice. - [Anatomy of an autonomous support agent](https://www.useclad.ai/blog/autonomous-support-agent): How autonomous AI agents intake, triage, investigate, resolve, and escalate — end to end. - [How AI support systems build compounding knowledge](https://www.useclad.ai/blog/compounding-knowledge): How AI support platforms build persistent knowledge models that compound institutional intelligence with each interaction. - [Your AI support tool has a comprehension problem](https://www.useclad.ai/blog/ai-comprehension-problem): The gap between AI that types and AI that truly understands — and why root-cause investigation matters. - [The architecture behind intelligent support triage](https://www.useclad.ai/blog/architecture-intelligent-triage): How AI triage systems correlate signals across channels, customer data, and knowledge bases to auto-classify and route support tickets in real time. - [AI support tools: what they are, how they work, and how to choose one](https://www.useclad.ai/blog/ai-support-tools): A comprehensive guide to AI support tools, how they work, and what to look for. - [From tickets to signals: how account intelligence changes support](https://www.useclad.ai/blog/account-intelligence): How support conversations contain the signals to predict churn, spot upsell opportunities, and keep accounts healthy. - [AI will reshape customer experience by 2030](https://www.useclad.ai/blog/ai-reshape-cx-2030): How AI will handle 80% of customer inquiries end-to-end within five years. - [Build vs. buy: the real cost of rolling your own AI support agent](https://www.useclad.ai/blog/build-vs-buy-ai): Why the hidden costs of building your own AI agent add up faster than most teams expect. - [How AI gives startup support teams an unfair advantage](https://www.useclad.ai/blog/ai-startup-advantage): How a 3-person team delivers the service of 30 with the right AI tools. - [Why cross-functional collaboration is the foundation of great support](https://www.useclad.ai/blog/cross-functional-collaboration): How shared context between support, success, and engineering leads to faster answers. - [Designing for trust: five principles for responsible AI in support](https://www.useclad.ai/blog/designing-for-trust): Five pillars that make AI trustworthy in customer support. - [The AI trust gap: security, privacy, and control in customer support](https://www.useclad.ai/blog/ai-trust-gap): How to close the gap between fast AI-powered support and customer data safety. - [The omnichannel playbook: unify every support channel](https://www.useclad.ai/blog/omnichannel-playbook): How to unify email, Slack, chat, and every other channel into a single support workflow. - [How to test AI agents before they talk to your customers](https://www.useclad.ai/blog/testing-ai-agents): Why multi-turn testing with noise injection is the only way to know if your AI agent is production-ready. - [Incident management for support teams: a practical framework](https://www.useclad.ai/blog/incident-management-framework): A step-by-step framework for detecting, triaging, resolving, and learning from incidents. - [Why UX research is the missing piece in most CX strategies](https://www.useclad.ai/blog/ux-research-cx-strategy): Why CSAT and NPS are not enough and how UX research improves support strategy. ## Company - **Name:** Clad (Clad Labs, Inc.) - **Website:** https://www.useclad.ai - **App:** https://app.useclad.ai - **Docs:** https://docs.useclad.ai - **Backed by:** Y Combinator - **Category:** AI-native customer support platform - **Tagline:** Support that resolves itself. - **Contact:** support@useclad.ai - **LinkedIn:** https://www.linkedin.com/company/clad-labs ## Getting Started - [Book a Demo](https://www.useclad.ai/get-started): See Clad in action on your own stack. - [Features Overview](https://www.useclad.ai/features): Walkthrough of triage, investigation, resolution, memory, and control. - [Tools Overview](https://www.useclad.ai/tools): Integrations, help center, widget, and API. - [About Clad](https://www.useclad.ai/about): The story behind the company. - [Sign In](https://app.useclad.ai): Access the Clad platform. ## Legal - [Terms of Service](https://www.useclad.ai/terms) - [Privacy Policy](https://www.useclad.ai/privacy) - [Subprocessors](https://www.useclad.ai/subprocessors)