Service Level Agreement (SLA)
Ref. connect.speechi.com – Speechi Connect
Last update date: June 28, 2024
Introduction
The purpose of this document is to define and clarify the commitments made by Speechi to its Users, through Service Level Agreements (SLAs), which are essential for using our services under the best possible conditions. By committing to high standards of reliability and performance, Speechi Connect demonstrates its commitment to excellence and user satisfaction. These SLAs reflect our commitment to providing an uninterrupted, high-quality user experience, guaranteeing the availability, performance and support needed to meet our users’ requirements. By detailing the measures taken to ensure the smooth operation of the Speechi Connect platform, this document highlights our determination to build a relationship of trust and support the success of our users in an ever-changing digital environment.
Availability
Service availability
Speechi guarantees the availability of its services 24 hours a day, 7 days a week.
Availability (real availability ): 99.5%.
The service is considered available when the methods create a response to a defined request. Service availability is measured at the processing center’s Internet transfer point. The service is considered unavailable when one or more interface methods fail to respond. This value is defined as a percentage of availability over a one-month period, 24 hours a day, 7 days a week. This method of calculation does not take into account maintenance time. Monthly availability is calculated as follows:
| Total VB availability (minutes) | No. of days/month x 24 hours x 60 minutes |
|---|---|
| Maintenance WG (minutes) | Maintenance time (minutes) |
| VN net availability (minutes) | VN = VB – WG |
| Outage time/panel | A |
| Actual availability per month | V = (VN – A) / VN |
In the following text, “service availability” means actual availability.
Availability measurement
Service availability is measured according to the following time intervals:
| Measuring service availability | 5 minutes |
|---|
Availability is measured automatically according to the interface methods (24h x 7d) at a defined time interval (every 5 minutes). Results are saved for analysis. Availability is measured at the data center’s Internet transfer point.
Maintenance
Service maintenance
The maintenance phase is a period during which services are not accessible or available. It enables us to check and improve the service provided to users.
Maintenance schedule
Maintenance schedules are scheduled according to the agenda below:
| Time slot for maintenance service with interruption < 10 min | Tuesday, 8:00 pm – 6:00 am |
|---|---|
| Time slot for maintenance service with interruption > 10 min | Wednesday 00h00 to 06h00 |
| Advance notice (Maintenance with interruption > 10 min.) | 3 working days |
A monthly framework is set for the maintenance service. WG maintenance time during maintenance slots, which may impact Speechi Connect platform performance, will be deducted from total VB availability and therefore will not be considered as a lack of SLA performance. Impacts on performance outside maintenance time slots will be included in the calculation of measured values for SLA.
For maintenance operations lasting more than 10 minutes, users will be notified three days in advance. This notification will include the following points:
- Maintenance type
- Systems affected
- Date and time of intervention
- Maintenance time
Reaction time
Defining reaction times
Reaction times are defined as follows:
- Reaction time: This is the time lapse between detection of the problem and the start of its resolution.
- Category A error – blocking fault: Service considerably reduced for the end customer, the user cannot use all or part of the solution.
- Category B error – major fault: Slightly reduced service for the end customer, the user can use the solution and the faulty functions can be performed via a workaround.
- Category C error – minor anomaly: Minor operational disruption, with no impact on the end customer. The user can use the solution without any blocking or workarounds.
Our reaction times
If an error occurs, either through direct detection by Speechi’s technical teams or through an error message from a user, intervention to resolve the problem must be implemented within a well-defined timeframe – this is the reaction time. Incident reports should be sent to Speechi by telephone or e-mail.
Reaction times vary according to the severity of the problem and the time of day it occurs:
| Error category | Days of the week | Opening hours | Reaction time | Message Status |
|---|---|---|---|---|
| Category A | Monday to Friday | 08 :00 – 18 :00 | 15 minutes | Every hour |
| 18 :00 – 08 :00 | 120 minutes | |||
| Weekends and public holidays | 00 :00 – 24 :00 | 120 minutes | ||
| Category B | Monday to Friday | 08 :00 – 18 :00 | 60 minutes | Every day |
| Category C | Monday to Friday | 08 :00 – 18 :00 | < 1 week | Every week |
Category A errors will have absolute priority and will be dealt with immediately by Speechi.
Compliance with SLA terms
Report
Speechi draws up regular reports to notify violations of the terms of the agreement.
- Availability and performance report as well as total service availability/performance per month.
- Incident statistics for all performances and services, broken down by service type and listing all problem types, resolution report, start time, end time and duration of disruption.
- Overview of violations of SLA terms and a description of the corresponding problem.
This report is available on request from support@speechi.com
SLA adjustment
Speechi reserves the right to modify the Service Level Agreement (SLA) at its discretion. Users will be notified of any significant changes in a timely manner.
