Usage
- 18 Oct 2023
- 5 Minutes to read
- Contributors
- Print
- PDF
Usage
- Updated on 18 Oct 2023
- 5 Minutes to read
- Contributors
- Print
- PDF
Article summary
Did you find this summary helpful?
Thank you for your feedback
| Usage | Description |
|---|---|
| User management | CA admin can create, edit, or deactivate users, define roles with specific permissions, and assign them to users via Identity Management. The admin can also create, edit, or remove departments. Departments are used to isolate data, allowing each user to interact with relevant information. It is important to note that all users must have at least one role that defines their permissions on the product, and each user must be assigned to a department. Another option for user management is the API, which includes several methods to create and edit users, roles, and departments. |
| Conversations | CA user can view all conversations and filter them based on their dates and conversation metadata properties. The filters are managed according to the available conversation types. For instance the conversation list can be filtered according to caller/called number, analysis status, listening status, min hold duration or Pbx agent ID only if the tenant has Conversation type call feature enabled. Similarly, if the chat feature is enabled the list can be filtered by considering title or participant. Acoustical analysis and speech-to-text results, details about categories and attached data can be viewed on the corresponding conversation's page. Users can provide comments, assign tags, or download the audio file related to a conversation from its detail page. For further analysis of the conversation list, users have the option to export it in xlsx format. |
| Categories | Categories are the foundation of almost every feature in CA. Detailed categories can be generated by adding logical operators and used to filter conversation groups. They can be created by user or AI generated both are managed on categories page: 1. User generated The one must select the conversation type and category language to succesfully and efficiently categorize the conversations. * If the category language has been selected they can be matched with the conversations with the conversations with the same language. * The conversations with no language selected can only matched with the categories enabled to work on all languages. |
| Root-cause analysis | CA enables the user to conduct more detailed statistical analysis. By creating a conversation group, users can observe frequently used phrases and detailed acoustical parameters. Additionally, it allows the user to create two separate conversation groups and compare them. CA can display the statistical behavior of categories over daily, weekly, or monthly periods. Non-FCR marking supports not only call-type conversations but also chats or video-type interactions. CA can analyze calls by considering the customer's phone number instead of the unique customer key. However, a unique customer key must be included in the conversation metadata for non-FCR tagging in other conversation types. There are three options to detect non-FCR conversations: 1. Only time settings: It is only required to satisfy the rules in the root cause settings to enable system to match conversations as non-FCR 2. Common user defined non-FCR category: It is not enough to satisfy settings. It is also required to have non-FCR category matched with the conversations. 3. Advanced non-FCR: It is not enough to satisfy settings. It is also required to have common Categories matched with the conversations. Meaning that the conversations should have the same topic in order to relate them. |
| AQM | Conducting QM includes creating forms, issuing evaluations from conversation details, or creating assignments to manage better evaluator workload and create automatic assignments for question types that can be easily evaluated by CA. An evaluation form can include both automatic questions (marked by CA) and manual questions (marked by the evaluator). The form creator can insert rules, normalize form scores, define reasons for questions, and review the form before publishing it. After creating a form, the evaluator can directly select a conversation from the conversations screen and evaluate it using the corresponding form. Another way to conduct evaluations is creating assignments. Assignments help organizations selecting more relevant conversations instead of using haphazard sampling by filtering conversations with categories. Assignments can be created and assigned to multiple evaluators. The evaluation performance of the evaluators and the performance of the agents are shown on the AQM dashboard and Evaluation Results page. Automatic assignments can also be created using forms that include only automatic questions. Automatic assignments evaluate every conversation in the selected department or agent list. Their results can be viewed on the Evaluation Results page and reflected on the dashboards. |
| Real-time notification | Real-time analytics is an online notification tool that informs selected users for defined triggers, such as abnormal agent speech rate, abnormal agent monotonicity, customer negative sentiment, or the detection of certain phrases. The users are required to create scenarios by selecting the users to monitor and setting triggers such as user-defined words, phrases, agent monotonicity, and more. Additionally, the users can define the notification content and select the recipients to receive the notifications. |
| Agent training | The training feature covers the practice and testing processes of a new agent. The trainer marks an analyzed call as training data and includes it in the training processes. During the training process, the agent practices the required transcript to be said. The correct script appears on the screen in practice mode, unlike the test mode. The SR system scores the agent's performance by comparing the exact script with the agent's trial. |
| Reports | CA can be integrated with BI tools such as Tableau and Power BI. Reports created in the BI tools can be represented on CA. In addition to Premium Reporting, CA offers an email reporting feature used to notify the users about conversations detected by a category. The user can create an email report by selecting the reporting period (immediate, hourly, or daily), date range, category name, and recipient email information. Immediate email reports instantly notify the user about detected conversations that match the selected category. Hourly and daily reports inform the user about conversations that fit the selected category within the last hour or day. |
| Logging | Audit Logs: An audit log object is typically created & saved per web request. They available under audit-logs page. It includes; -Request & response details (like URL, Http method, Browser info, HTTP status code... etc.). -Performed actions (controller actions and application service method calls with their parameters). -Entity changes occurred in the web request. -Exception information (if there was an error while executing the request). -Request duration (to measure the performance of the application). Security Logs: A security log object is typically created & saved per identity server request, including time, action, ip address, browser, username details.They are available under identity/security-logs page. Action types: Login succeeded, Login failed, Login not allowed, Login requires 2FA, Logout Impersonate, change email and change password details. Action Logs: An action log object is created & saved per web request. It includes the conversation related event, its details and the occurance time. the events are: conversation's analysis progress items such as analysis, waveform generation, media reording, categorization and automaic evaluation. Additionaly listening status with the listener's username and the evaluation history are presented in conversation's details under history tab. |
Was this article helpful?