Customer support
| Plan | Free | Basic | Professional | Expert |
|---|---|---|---|---|
| Help documentation | Self-service onboarding | Self-service onboarding | Assisted onboarding | Assisted onboarding |
| Website live support | 24/7 | 24/7 | 24/7 | 24/7 |
| Onboarding guidance | Initial AI setup assistance | Full product-configuration support | ||
| Dedicated support group | Weekdays: 9:30 a.m.–6:30 p.m. UTC+8 | 24/7 priority support |
Issue priorities
Priority 1 (High)
Priority 1 covers major system failures or unavailable core functions that prevent critical customer work, or an active data leak or data loss.
Examples include:
1. Lost conversation messages
2. Unable to sign in to the workspace
3. Workspace stuck loading or showing a blank screen
Priority 2 (Medium)
Priority 2 covers recurring disruption caused by malfunctioning product functionality, or a risk of data leakage or loss.
Examples include:
1. Conversation-origin page unavailable
2. AI unable to respond normally
3. Delayed messages
Priority 3 (Low)
Priority 3 covers issues with limited impact on daily work, rarely used functionality, or a low risk of data leakage or loss.
Examples include:
1. UI display issues
2. Problems on pages not critical to conversations
3. Configured settings not taking effect
Service-level targets
Priority 1: Initial response within 30 minutes of submission and remedial action within 60 minutes, with continuous 24-hour support.
Priority 2: Initial response within one hour of submission and remedial action by the next business day.
Priority 3: Initial response within two hours of submission and remedial action within the next two business days.
Resolution process
Remedial action may be a software fix, workaround, or action plan.
An action plan includes:
1. Confirm whether the reported behavior is a defect or intended product behavior.
2. Reproduce the issue. If reproduction fails, clarify the conditions and verify again.
3. Collect details for engineering, such as workspace ID, conversation ID, visitor identifier, time observed, and reproduction steps.
4. Engineering reviews databases, logs, and code to identify the cause.
5. Evaluate possible solutions and consult the product team when needed.
6. Update code, tests, or configuration.
7. Test in an internal environment.
8. Deploy after QA approval.
9. Ask the user to verify. Return to step 1 if the issue remains.
These targets are intended to minimize unexpected downtime and support business continuity.
Release cycles
Requests and issues reported during a release cycle are assessed or addressed in the following cycle.
CDN setup
Mixdesk uses AWS CloudFront and Cloudflare. During a CDN incident, a DNS CNAME change can switch the CDN.
CDNs normally refresh from the origin automatically. Mixdesk can also trigger a refresh through the CDN management console when needed.
Monitoring and incident response
Mixdesk’s operations team uses Prometheus and AWS CloudWatch, with dashboards and alerts for business metrics. Engineering and operations teams follow up on incidents.