Features
- 10 May 2024
- 6 Minutes to read
- Contributors
- Print
- PDF
Features
- Updated on 10 May 2024
- 6 Minutes to read
- Contributors
- Print
- PDF
Article summary
Did you find this summary helpful?
Thank you for your feedback
| Feature | Description |
|---|---|
| On-cloud and on-premise availability | Providing clients with the flexibility of deploying our solutions both on-cloud and on-premise, ensuring accessibility in any preferred environment |
| Rich language support | Virtual Agent languages: Turkish, English, Arabic, Urdu, Russian, Latvian, Spanish, French, Dutch, German and Italian UI languages: Turkish, English and Arabic |
| Low-code business flow design | - Flexible and customized scenario flow design according to the needs - The flow design is visually based to create an easier experience - Supports rich content outputs: quick reply, location, card, carousel, dynamic buttons - Predefined and reusable contents such as virtual agents, entities, flows - Testing the scenario on the flow design screen - Flow designs include nodes which are functional elements connected by links. Nodes’ features are for instance: >Ask a question or a missing slot >Detect the intent, extract the entity >Define the condition and route the flow based on the condition >Send HTTP request >Nodes for call transfer, and live agent handover |
| Machine-learning-based state of art natural-language-understanding | - Text-normalization features to apply spell correction - On-demand testing of utterance - intent matching - Learning from fallback utterances |
| Benefiting from Open AI | - Utterance suggestion - Conversation summarization - Transferring summary to the Live Agent -Creating Project - Alternative Response Text - FAQ Utterance Multiplication - Gen QnA Feature |
| Integrations with a variety of channels and services | Channel integrations are supported by a variety of providers: Unifonic, Infobip, Twilio,Microsoft Bot. Our supported channels integrations are with: Avatar Slack SMS Teams Telegram Webchat IVR Service integrations encompasses integrations with our products or solutions. Our service integrations are: Text-to-Speech Speech Recognition Voice Biometrics Live Chat Sentiment Analysis |
| Dashboard for overviewing the statistics of the conversations | The dashboard includes graphics for: Sessions: Session graphs help to understand session traffic. There are graphs for: Total Sessions Average Session Duration Average Message per Session Intents: Intent graphs provide insights about users' transactions. Top Intents: The most common intents among users Intent Overview: Matched/unmatched number of intents Intent Trends: Number of intents over time Fallback Messages: Messages that do not match an intent Sentiments: Conversation-based sentiment scoring techniques track user satisfaction in sessions. Average Sentiment: Ranges from very positive to very negative Sentiment Trends: Sentiment scores over time Average Rating: Users' own rating at the end of the session |
| Conversations page to see users' conversations firsthand | - Filtering conversations according to the project, channel, and dates - Being able to search with conversation/session ID - The conversation details page shows the dialog between the customer and the chatbot/live agent. This page also indicates specific info about one session such as: > Customer information > Customer sentiment > Summary of the conversation |
| Component | Description |
|---|---|
| Project | The overall design of the end user's end-to-end conversation experience is called the Project. It contains main and sub-flows created for the workflow to be followed. |
| Virtual Agent (VA) | The component using the Natural Language Understanding (NLU) unit to interpret the utterances of the users to match them with the correct intents is known as the Virtual Agent. In the created virtual agent, the utterances are labeled with the associated intents and trained accordingly. The project uses the Virtual Agent to explain and perform the operation. For example, the “deposit money” intent can be created inside the Banking virtual agent. One of the utterances of this intention can be “deposit money”. By training the virtual agent with these utterances, when the user says, "I want to deposit money into my savings account", it matches this expression with the “deposit money” intent. To give another example, when the user says, “I lost my debit card”, the trained virtual agent matches this statement with the “lost/stolen card” intent. Adding utterances that are close to the user’s statements, helps the virtual agent to make correct interpretations. Virtual agents can be saved in a pre-defined form and they can be connected easily to the projects to be created later. Also, the designer can easily use the virtual agents to make changes according to the customer’s request. |
| Utterance | A spoken/written word, statement, or vocal sound given by the customer to explain their request is called an utterance. For example, if the end-user uses the phrase "show my account balance", their phrase is considered an utterance. |
| End-user | The user who specifies their needs to the system with their utterances is described as the End-user. |
| Tenant | Tenants allow users to manage multiple projects from a single server. Each tenant is unique, and changes to one do not affect the others. Users can log in to their tenant with a predetermined user code. |
| Intent | Intent is the actual action that the user wants to take. An utterance given by the user is matched with the intent having the highest confidence value. Also, the intent contains multiple utterances to match the customer’s newly given utterance with the correct intent. For example, the actual intent of the expression ‘I lost my credit card’ is card cancellation. Proper action is taken by the system after understanding the intent of the requester. |
| Entity | Entities are variables used for extracting, mapping, and delivering operation or user-based dynamic data. For example, if a user types “show me my saving account balance for January”, the entities are specified as “saving account” and “January”. Entities are given a name such as “Account Type” and “Month,” and they are sometimes referred to as slots. To give another example, the entity [cardtype] in this expression “I want to apply for a Gold credit card.” is Gold. Customers' National IDs or phone numbers can be given as the final examples of the entities. |
| Action | The utterance matching the intent with the highest confidence value performs the relevant operations within the framework of the action to be taken by this intent. Flows in the actions are created with nodes with different functions. Canceling the customer’s credit card can be given as an example of the action. |
| Node | The designs in the project and the virtual assistant are made through structures with different functions called "nodes". For instance, the node is used to ask questions to the customer or to manage the transactions to be made. |
| Confidence Value (CV) | Confidence value is a value between 0-1 that indicates the confidence of utterance match the correct intent name. |
| Integration | Integration technically enables software and systems to communicate and to exchange data in an automated manner. Projects can be associated with channels by adding integration configurations. Thanks to these integrations, the product can be reached by customers via different channels, and a project can have multiple channel integrations. For instance, in order to provide a service to the customer through WhatsApp channel, the project must have a WhatsApp integration. |
| Flow | Flows are created by the scenario designer, and they start working when the end user sends a request to the system. The actions to be taken are organized with the scenario flow created by the nodes added between the start and end nodes. In the main flow, it is determined how the end user will continue their transactions after logging into the system. Also, this can be supported by sub-flows that will be added to the main flow. For example, when the chat ends, the survey directed to the end user can be designed under a different sub-flow and added to the main flow as a node named a sub-flow. Creating and saving these sub-flows separately and adding them to the main flow provides ease of use to the scenario designer and simplicity to the main flow. |
Was this article helpful?