This article discusses AI Commerce platform’s service level agreements (SLAs) and support failure terms. It gives you a clear picture of how AI Commerce ensures near-constant availability, what priorities are set for support requests, and how to respond to potential outages and other issues. This will help you understand how the SLAs apply in practice and what rights and obligations they entail.

AI Commerce 's SLA and its coverage

• These terms apply only to AI Commerce platform's own SLAs. Agreements with potential integration partners are outside the scope of this article.
• AI Commerce offers 99.999 percent uptime, meaning services are virtually always available.
• High availability of services means that pages respond to requests and functions critical to commerce (such as logging in and ordering) are not blocked.
• Limited application errors or partial problems (for example, the lack of one payment method) are not considered an outage, but have their own support service agreement level.
• The highest availability tracking is displayed on the public Page Status page, where you can also see historical availability information.

Support service priority levels

AI Commerce support requests are classified into four different priorities, each of which defines response and repair times and the urgency of actions.

1. P1 - critical:
   The entire store or essential function is completely down (for example, purchasing is completely blocked). Response time less than 1 hour (24/7 on-call). Repairs are started immediately and continue without interruption until the problem is resolved. Additional resources are alerted if necessary. The customer is informed of the progress of the error investigation and repair at least every hour.
2. P2 - high:
   A serious disruption that significantly disrupts business operations (e.g. critical functionality is down but does not bring down the entire service). Response time typically within 6 hours during business hours. Repairs begin immediately during business hours and continue until the fault is resolved or bypassed.
3. P3 - normal:
   Some functionality is disrupted, but a workaround exists or the impact is limited. Response time is 24 hours (no later than the next business day). These bugs will be fixed within the normal release cycle, for example in the next version update.
4. P4 - low:
   A minor, non-critical issue (e.g., cosmetic bug, ERP interface bug, or general question that is not urgent). Response time 1–3 business days. This also includes development requests that do not require an immediate fix and are implemented within the normal product development schedule or advice is provided without a specific timeline. 
    

Service availability and maintenance outages

• If the availability of services falls below 99.999%, AI Commerce will reimburse 10% of the monthly fee for each percentage point.
• Downtime means significant or complete disruption (for example, products cannot be moved to the shopping cart or no payment method can be used at checkout).
• A partial or minor malfunction (e.g., customer prices not being visible) is generally not a direct outage because trading is not blocked. The customer can place an order, and the customer prices will be adjusted for the order placed.
• AI Commerce reserves the right to perform maintenance outages during the night (typically 00:00–07:00). Major maintenance outages will be announced at least 48 hours in advance. Very short outages may not be announced in advance, but will be scheduled for the quietest time (approximately 02:00–03:00).
• Typically, P3 and P4 support is provided only during business hours on weekdays. Updates provided on weekends are not covered by the SLA, but are additional services.
• P4 category cases (user error, cosmetic bug, development request) are not covered by liability, especially if the problem is resolved through a self-service channel.

AWS service exceptions (Force Majeure)

• AI Commerce leverages two different AWS Availability Zones. If AWS is unable to deliver services in either environment, AI Commerce will notify customers.
• AWS is liable for damages according to its own SLAs. AI Commerce is not directly responsible for these, but will keep customers informed.
• AWS is a very reliable provider, and service outages are very rare.

Refunds and sanction

• If the service outage exceeds the allowed limit, AI Commerce will refund 10% of the monthly fee for each percentage point.
• AI Commerce does not compensate for actual revenue or other losses, as this is impossible to measure. There is a separate information article on this topic.
• Other forms of compensation are offered for customer service failures (for example, consultation or prioritization in the development queue).
• AI Commerce is not liable for compensation in cases where the problem could have been solved with AI-assisted self-service. This encourages the use of self-service channels first, and allows customers to have better experts and quality in human customer service. 
   

Summary

AI Commerce 's service level agreement (S-segment) guarantees high (99.999/59%) availability, fast response to critical issues, and clear escalation to other support requests. Downtime is defined as a significant factor preventing trading, while minor disruptions fall into other priority categories. If the service level is not achieved, AI Commerce offers various compensations, but strongly encourages the use of artificial intelligence and self-service channels. This ensures that the support team's resources are allocated as efficiently as possible to handle the most critical issues.