How long does it take to set up RAUM AI?

Most teams are fully operational within a single afternoon. Install the Zendesk marketplace app, connect your API key, and RAUM AI starts learning your ticket patterns immediately.

Does RAUM AI replace human agents?

No. RAUM AI augments your team by handling repetitive tickets autonomously, so your agents can focus on high-value, complex interactions that require a human touch.

What AI models does RAUM AI use?

RAUM AI is BYOK (Bring Your Own Key) — you connect your own OpenAI, Anthropic, or Google Gemini API key. You control the model, the cost, and the data. We never mark up AI provider fees.

Is my customer data safe?

Absolutely. We employ PII redaction before any data reaches AI models, use AWS KMS encryption for all credentials, and maintain strict multi-tenant data isolation.

Can I customize the AI's tone and behavior?

Yes. You can configure guidelines, decision trees, knowledge bases, and brand voice settings per organization. The AI follows your exact instructions.

What happens when I exceed my auto-resolution limit?

You can choose: either pause auto-resolutions until next month and get notified, or continue resolving with a small per-ticket overage fee. This is configurable per organization.

Is copilot usage (agent assists) really unlimited?

Yes. Every plan includes unlimited AI copilot usage for your agents — no caps, no metering, no surprise charges. We only meter automated resolutions where AI resolves tickets without human involvement.

The AI Copilot That Thinks Before It Replies

AI that resolves tickets,
not just replies

A multi-agent reasoning engine running natively inside Zendesk. Four stages — Understand, Plan, Act, Reply — for every single ticket. Reduce support costs by 80% with your own AI keys.

Install on Zendesk See how it works

AI provider markup — pay at cost

4-stage

Multi-agent reasoning pipeline

<5min

Setup time in Zendesk

∞

Copilot usage on every plan

Under the hood

A reasoning engine,
not just a prompt wrapper

Most AI copilots fire a single prompt and hope for the best. RAUM routes every ticket through a four-stage reasoning chain before generating a single word.

🧠

Understand Stage 1

Reads the ticket, analyzes intent, extracts context from Zendesk tags and conversation history.

📋

Plan Stage 2

Searches your decision trees and knowledge base, maps out the right steps to solve it.

⚙️

Act Stage 3

Executes MCP tools, fetches live data from your APIs, and takes autonomous action when configured.

✉️

Reply Stage 4

Writes a response in your brand voice — streamed in real-time directly into the Zendesk composer.

raum-pipeline · ticket #18472

// Understand — intent analysis

intent: "billing_dispute",

confidence: 0.94,

tags: ["vip", "billing"]

// Plan — resolution path

tools: ["billing_api", "kb_search"],

escalate: false,

auto_resolve: true

// Reply — streaming response

"Hi Sarah, I've reviewed your invoice..."

How Teams Use RAUM

Built for teams of every size

🚀

Small Teams (1–10 agents)

Automate repetitive tickets from day one. Get back 20+ hours per week without hiring. Setup in under 5 minutes — no dedicated AI team required.

📊

Mid-Market (10–50 agents)

Scale support without scaling headcount. Track resolution rates and costs with advanced analytics. Configure MCP tools for your stack and maintain consistent quality.

🏢

High Volume (50+ agents)

Custom infrastructure, dedicated onboarding, and SLA agreements. PII redaction and KMS encryption keep your data compliant. Knowledge gap detection surfaces what's missing.

What you get

Everything your support team needs.
Nothing they don't.

Two modes, one platform. Deploy RAUM as a fully autonomous agent or give your human team an intelligent copilot — or both, simultaneously.

✨

AI Reply Generation

Drafts complete, context-aware ticket responses based on conversation history, Zendesk tags, and your knowledge base. Streamed in real-time into the Zendesk composer.

Copilot Mode

🤖

Autonomous Resolution

Webhook-driven auto-resolution closes tickets without human intervention. Your team only sees what couldn't be handled automatically.

Agent Mode

📊

Analytics Dashboard

Track cost per ticket, resolution rates, token usage, and agent vs copilot breakdowns. Activity heatmaps, CSAT tracking, and knowledge gap detection.

All Plans

🎨

Brand Voice Control

Configure tone, style guidelines, and decision trees per organization. The AI follows your exact instructions — not generic defaults.

All Plans

🔧

Tool Integrations

Connect external APIs and services. Run live lookups, trigger actions, and integrate with your existing stack easily.

Copilot + Agent

🔑

Bring Your Own Key

Use your own OpenAI, Anthropic, or Gemini API key. Your data never passes through our AI accounts — full cost control and data sovereignty, built in by design.

Enterprise-ready

🌍

Grammar Fix & Translate

Automatically detects customer language, translates your reply, or fixes grammar before you hit send.

Copilot Mode

⚡

Response Cache

Zero-latency instant replies for repeated queries. Skips AI processing to reduce cost and resolution time.

All Plans

🧪

Simulation Mode

Dry-run the full AI pipeline against real tickets without executing actions. Test and configure with confidence.

Admin Tools

Pricing

Simple, transparent pricing.
One variable: ticket volume.

No per-seat AI charges. No hidden copilot metering. You bring your AI key — we charge only for the platform.

Starter

$79/mo

For small teams getting started with AI support automation.

Save 16% billed annually at $790

Up to 5 agents
200 auto-resolutions/mo
Unlimited copilot usage
Unlimited AI tools & languages
Unlimited ticket analysis
Unlimited tool integrations

Start Free Trial

+$0.20 per extra auto-resolution

What early users are saying

★★★★★

"The autonomous resolution mode alone justified the cost in the first week. We went from handling 400 tickets a day to our team only touching the ones that genuinely need a human."

Marcus K.

Head of Support · SaaS company, 18 agents

★★★★★

"Finally an AI copilot that actually understands our product. The multi-agent reasoning means it doesn't just hallucinate an answer — it actually looks up the right procedure before replying."

Anika L.

Support Lead · E-commerce, 9 agents

★★★★★

"The BYOK model was the dealbreaker for our security team. Our ticket data never leaves our API key's account. Setup was genuinely under 10 minutes from the Zendesk Marketplace."

Tom R.

CTO · FinTech startup, 30 agents

Ready to resolve tickets
while you sleep?

14-day free trial. No credit card surprises. Cancel anytime. Your Zendesk setup, supercharged in under 5 minutes.

Install on Zendesk Book a demo

AI that resolves tickets,not just replies

A reasoning engine,not just a prompt wrapper

Understand Stage 1

Plan Stage 2

Act Stage 3

Reply Stage 4

Built for teams of every size

Small Teams (1–10 agents)

Mid-Market (10–50 agents)

High Volume (50+ agents)

Everything your support team needs.Nothing they don't.

AI Reply Generation

Autonomous Resolution

Analytics Dashboard

Brand Voice Control

Tool Integrations

Bring Your Own Key

Grammar Fix & Translate

Response Cache

Simulation Mode

Simple, transparent pricing.One variable: ticket volume.

What early users are saying

Ready to resolve ticketswhile you sleep?

AI that resolves tickets,
not just replies

A reasoning engine,
not just a prompt wrapper

Everything your support team needs.
Nothing they don't.

Simple, transparent pricing.
One variable: ticket volume.

Ready to resolve tickets
while you sleep?