50% resolution in 30 days. Or every dollar back. See the guarantee →

Pricing

A 20¢ ticket shouldn't cost $1.

Decagon, Sierra, Maven, Gorgias, and Zendesk charge a flat ~$1+ per resolution. Same price whether the AI wrote one sentence or ran a 20-turn refund flow. We meter by the token: simple tickets can be as low as $0.10, complex ones cost more, and you see every cent on the bill.

Pro
2 seats included · 833 credits · ~1,500 conversations
Does the work of ~1 human agent
~$0.35/conv
$500/mo
Pro Max
5 seats included · 4,500 credits · ~7,500 conversations
Does the work of ~5 human agents
~$0.32/conv
$2,500/mo
Enterprise
10 seats included · 11,000 credits · ~19,000 conversations
Does the work of ~15 human agents
~$0.29/conv
$5,500/mo

Conversation estimates use the model you selected above. Varies by complexity. Try our pricing calculator →

Customize your plan: enter your numbers
Build your plan

See exactly what you'll pay.

AI models. Each has different speed, intelligence, and credit cost

Loading models…

Credit estimates are approximate. Actual usage varies by conversation length and complexity. You only pay for what you actually use.

Need more? Extra seat: $100/mo. Credit packs:

1K
$600
$0.60/cr
5K
$2,750
$0.55/cr
-8%
10K
$5,000
$0.50/cr
-17%
25K
$12,500
$0.50/cr
-17%
Demo call → Week 1 live → Day 30: 50% resolved or full refund

Switching from Gorgias or Zendesk? We migrate 30 to 50 brands per month from these platforms. One-click import for tickets, macros, tags, and users. Everything comes with you.

Savings Calculator

Calculate how much you save.

Hourly payroll converts using 8 hours x 22 days.
Conservative 40%Optimistic 85%
Human hires
$0
Current team cost
AI + retained
$0
Richpanel + retained team
Savings
$0
Per month
Annual
$0
Per year

Brands using Richpanel AI see 50%+ resolution rates within 30 days, with CSAT scores above 4.5.

Why CX leaders trust us

Built for teams that can't afford to get it wrong.

Three non-negotiables for any AI touching your customers: it can't make things up, it can't leak data, and it can't break under load. Engineered, audited, and load-tested before it ever ships to your inbox.

Hallucination-proof.

Grounded in your data. Never invents information. Escalates the moment confidence drops below your threshold.

Cited sources Auto-escalation 1,000+ simulations on your SOPs

Enterprise-grade security.

Your data never trains models. EU-friendly, audited, and built to pass procurement on day one.

SOC 2 HIPAA GDPR CCPA

Infinitely scalable.

Black Friday surges and 3 AM queues. Scale up and down instantly without re-architecting anything.

3–5× peak 24 / 7 / 365

FAQ

What if it doesn't hit 50%?
You get a full refund. We set up on your real business, with your policies and integrations. If your AI hire doesn't resolve at least 50% of conversations autonomously in 30 days, we refund every dollar. No proration, no fine print.
How do you prevent hallucinations?
Four layers. First, pre-launch evals run 1,000+ simulations on your SOPs before the AI ever talks to a customer. Second, a separate QA AI reviews every reply before it sends. Third, actions like refunds and cancellations use deterministic tool execution, not guesswork. Fourth, when confidence drops, it escalates to a human instead of guessing. The result: grounded in your data, never invents information.
We tried AI chatbots before and they were terrible.
So did most of our customers. Old chatbots were decision trees pretending to be smart. Richpanel AI uses frontier LLMs (Claude, GPT) grounded in your actual data: orders, policies, history. It reasons, not pattern-matches. And when it's unsure, it escalates instead of guessing. Night and day difference.
How long does setup take?
Give us your URL. CX Manager generates your full setup in one conversation. Under 20 minutes on average.
Can I control what the AI is allowed to do?
Completely. You set which actions require human approval (refunds, cancellations, exchanges) and which the AI can execute autonomously (order status, FAQs, tracking). Your SOPs are the AI's operating manual.
Do you import historical ticket data from our current helpdesk?
Yes. Past tickets, contacts, custom fields, macros, and tags all transfer. Macros become AI SOPs. Tags train brand voice. We migrate 30 to 50 brands per month from these platforms. Most migrations go live in 3 to 7 days with zero downtime.
Which channels do you support?
Email, live chat, social messaging, WhatsApp, SMS, and marketplace channels. All unified in one inbox. Deep integrations with Shopify, Loop, Recharge, TikTok Shop, AfterShip, and more. If you need something we don't have, tell us.
What if we're not ready to go live until next quarter?
Start setup now, go live whenever. Your configuration, SOPs, brand voice, and integrations all persist. Setup work doesn't get thrown out. When you're ready to flip the switch, you're already trained.
How do seats and credits work?
Seats are for humans using the helpdesk. Credits are AI capacity for model usage. Each plan includes both. If your team grows, add seats at $100/mo. If conversation volume grows, buy credit packs without changing the rest of your setup.
What happens when credits run out?
Conversations route to human agents. Buy credit packs anytime (1K, 5K, 10K, or 25K) for instant AI capacity. Larger packs save up to 17%.
Do unused credits expire?
You get extra time to use unused credits after your plan ends. Monthly billing gets a 15-day grace period; yearly billing gets 90 days. After that, unused credits expire.
Is my data safe?
Never trains models. SOC 2 Type II, GDPR, CCPA. Encrypted everywhere.
Our volume is low, under 50 tickets a day. Does this make sense for us?
Honestly: ROI kicks in above roughly 3 agents or ~2,000 conversations/month. Below that, the $500/mo plan can still pay back through 24/7 coverage and nights/weekends, but book a call first. We'll tell you honestly if it's a fit or if you should wait.
Can we use our own API keys instead of your credits?
We don't offer BYO API keys. Credits don't just buy tokens. They pay for the full hire: orchestration, safety layers, compliance, escalation, and continuous tuning. We negotiate bulk commitments with every AI provider, and those volume rates are baked into your credit price. That's how the per-conversation cost stays lower than any alternative.

What every hire includes

Same AI, same training, every plan.

No tiered safety. No locked integrations. No "contact sales for security." Your AI hire on Pro gets the same intelligence, compliance, and setup help as Enterprise. The only difference is how many conversations it handles.

✓ Speaks in your brand voice, not a generic bot
✓ Works 24/7/365, no shifts or overtime
✓ Email, chat, social, SMS, WhatsApp in one inbox
✓ Manage multiple brands from one platform
✓ QA AI reviews every reply before sending
✓ Escalates to humans when unsure
✓ Saves subscriptions before cancelling
✓ Shopify, Loop, Recharge, TikTok Shop, AfterShip
✓ SOC 2 + GDPR + CCPA compliant
✓ Gets smarter with every correction

50% resolution in 30 days.

Or your money back. No hoops.

We set up your AI on your actual business, with your real policies and integrations. If it doesn't resolve at least 50% of conversations autonomously in 30 days, we refund every dollar. No proration. No "but the data."

See it built on your brand →

Built live on the call · 20 min · No credit card