CS Copilot · Powered by Claude AI
MailTriage reads every customer email, scores its emotion and urgency, and drafts a reply that follows your store policies. Your team just reviews and sends.
Order #ORD-98421 arrived broken, please make this right!
J&T tracking #ORD-99120 stuck for 3 days
Bulk order inquiry: 250 uniforms
Denim jacket size chart, size L
AI summary
Glass tumbler shattered and damaged a jacket. Customer has an unboxing video and wants a full refund or replacement.
Draft reply · follows 3-day return policy
< 10 sec
from incoming email to draft reply
1–5
automatic urgency score
100%
replies reviewed by a human
0
data used for model training
Built for the emails online stores get every day
URGENCY 5/5
Broken, defective or wrong items handled first, with replies that follow your return policy.
URGENCY 3/5
Tracking numbers pulled out automatically, with calm, clear status updates.
URGENCY 2/5
Sizes, variants and stock questions answered fast so shoppers keep buying.
OPPORTUNITY
Bulk orders and reseller inquiries flagged so no deal gets lost in the queue.
Features
Broken items and held refunds jump to the top. General questions and B2B offers wait their turn.
Drafts follow your return policy, unboxing-video rules and courier guarantees. The AI never promises anything outside them.
Review drafts, switch between formal and casual tone in one click, then approve. Nothing goes out without sign-off.
Order numbers, SPX or J&T tracking codes and product variants are pulled from the email automatically.
Flag upset customers before they leave a 1-star review.
Including slang, abbreviations and the typical tone of marketplace shoppers.
Integrations
Connect in minutes. No migration, no new email address, no changes for your customers.
Sign in with Google and MailTriage starts triaging new emails right away.
● Available
Use your own domain or hosting provider. If it supports IMAP, it works.
● Available
How it works
Add your store name, support hours, tone of voice, and return and shipping policies.
Each email gets a category, emotion score, urgency, summary and recommended action.
Agents approve ready-made drafts. Response time drops from hours to minutes.
Security
MailTriage runs on Anthropic Claude (Sonnet & Haiku), chosen for its accurate contextual reasoning and strict privacy standards.
Customer emails are never used to train models.
Answers grounded in your policies, not guesses.
Every reply is approved by an agent before it is sent.
Pricing
Pricing is based on your monthly email volume. Book a demo and we will send a quote within one business day.
Standard
For growing online stores with one support team.
From
Rp 150.000/month
Volume-based, billed monthly
Custom
For high-volume brands and multi-store teams.
Custom pricing
Tailored to your volume & workflow
No. Every draft waits for approval from your support agent.
You fill in the policy knowledge base once, and the AI always refers to it when drafting replies.
Yes. Data is processed under a zero-retention policy and never used for model training.
A 20-minute session: we run real emails from your store scenarios through the simulator and walk through integration.
Standard covers one Gmail, Google Workspace or IMAP inbox with urgency and emotion scoring, policy-grounded draft replies and human approval before anything is sent. The final monthly price depends on your email volume.
Standard is billed monthly based on your email volume. After your demo we send a quote for your exact volume within one business day.
Contract terms are confirmed in your quote. Custom plans for high-volume or multi-store teams are agreed individually during onboarding.
A 20-minute demo, no commitment. We will show triage and draft replies live on your real cases.