| Action |
Actions represent the tasks or questions that your assistant can help customers with. Each action has a beginning and an end, making up a conversation between the assistant and a customer. Learn more. |
Actions |
| Ask clarifying question |
A feature that enables the assistant to ask customers to clarify (disambiguate) their meaning when the assistant isn't sure what a user wants to do next. Learn more. |
Actions |
| Assistant |
Container for your actions, channels, and integrations. You add actions and at least one channel to an assistant, and then deploy the assistant when you are ready to start helping your customers. Learn more. |
Both |
| Change conversation topic |
A feature that gives the user the power to direct the conversation. It allows digressions and prevents customers from getting stuck in a dialog thread. They can switch topics whenever they choose. Learn more. |
Actions |
| Channel |
The location where your assistant interacts with your users, for example, over the phone, on a website, or in Slack. At least one channel is required for every assistant. Learn more. |
Both |
| Completion |
Measures how often within a time period that a user reaches the end step of an action. Learn more. |
Actions |
| Content |
The conversation logic and words that are used to respond to your customer. Content is required for every assistant. Learn more. |
Both |
| Content catalog |
A set of prebuilt intents that are categorized by subject, such as customer care. You can add these intents to your dialog and start using them immediately. Or you can edit them to complement other intents that you create. Learn more. |
Dialog |
| Context variable |
A variable that you can use to collect information during a conversation, and reference it later in the same conversation. For example, you might want to ask for the customer's name and then address the person by name later on. A context
variable is used by dialog. Learn more. |
Dialog |
| Dialog |
A component where you can build the conversation that your assistant has with your customers. For each defined intent, you can write the response your assistant returns. Learn more. |
Dialog |
| Digression |
A dialog feature that gives the user the power to direct the conversation. It prevents customers from getting stuck in a dialog thread; they can switch topics whenever they choose. Learn more. |
Dialog |
| Disambiguation |
A dialog feature that enables the assistant to ask customers to clarify their meaning when the assistant isn't sure what a user wants to do next. Learn more. |
Dialog |
| Entity |
Information in the user input that is related to the user's purpose. An intent represents the task that a user wants to do. An entity represents the object of that action. Learn more. |
Dialog |
| Environment |
You can group your work in separate containers that are called environments. Each environment contains its own content, channels, and extensions. Environments also have their own IDs, URLs, and service credentials that can be referenced
by external services. Each new assistant comes with two environments: the draft environment and the live environment. Your customers interact with assistants on the live environment and cannot interact with assistants on the draft environment.
The separation of these two environments allows you to build and iterate on your content separately from what your customers see. You do not want customers to stumble upon an incomplete action that leads them to a dead end. For Enterprise
plans, you can add up to three environments as a staging area to test your assistant before deployment. You can build content in the draft environment and test versions of your content in the extra environments. Learn more. |
Both |
| Escalation |
If your assistant is integrated with one of the supported service desk systems, you can build in logic that transfers, or escalates, the conversation to a person when necessary. Learn more. |
Both |
| Extension |
A custom extension is an integration with an external service. By calling an extension from an action, your assistant can send requests to the external service and receive response data it can use in the conversation. Learn more. |
Actions |
| Incompletion |
Reasons why an action is not completed by a user, including escalated to agent, started a new action, stuck on a step, or abandoned or ongoing. Learn more |
Actions |
| Integrations |
Add-ons to the end experience that help solve specific user problems, for example, connecting to a human agent or searching existing help content. Integrations are not required for an assistant, but they are recommended. Learn more. |
Both |
| Intent |
The goal that is expressed in the user input, such as answering a question or processing a bill payment. Learn more. |
Dialog |
| Message |
A single turn within a conversation that includes a single call to the /message API endpoint and its corresponding response. |
Both |
| Monthly active user (MAU) |
A single unique user who interacts with an assistant one or many times in a month. Learn more. |
Both |
| Preview |
Embeds your assistant in a chat window that is displayed on an IBM-branded web page. From the preview, you can test how a conversation flows through your assistant, from end to end. Learn more. |
Actions |
| Recognition |
Measurement of how many requests are being recognized by the assistant and routed into starting an action. Learn more. |
Actions |
| Response |
To create your assistant's response in an action step, you use the Assistant says section. This represents the text or speech that the assistant delivers to a user at a particular step. Depending on the step, you can add a complete
answer to a user's question or ask a follow-up question. Learn more. |
Actions |
| Response |
Logic that is defined in the Assistant responds section of a dialog node that determines how the assistant responds to the user. When the node's condition evaluates to true, the response is processed. The response can consist of
an answer, a follow-up question, a webhook that sends a programmatic request to an external service, or slots that represent pieces of information that you need the user to provide before the assistant can help. The dialog node response
is equivalent to a Then statement in If-Then-Else programming logic. |
Dialog |
| Slots |
A special set of fields that you can add to a dialog node that enable the assistant to collect necessary pieces of information from the customer. For example, the assistant can require a customer to provide valid date and location details
before it gets weather forecast information on the customer's behalf. Learn more. |
Dialog |
| Step |
A step that you add to an action represents a single interaction or exchange of information with a customer, a turn in the conversation. Learn more. |
Actions |
| System entity |
Prebuilt entities that recognize references to common things like dates and numbers. You can add these to your dialog and start using them immediately. Learn more. |
Dialog |
| Subaction |
An action that is called from another action is referred to as a subaction. Learn more. |
Actions |
| Try it out |
A chat window that you use to test as you build. For example, from the dialog skill's "Try it out" pane, you can mimic the behavior of a customer and enter a query to see how the assistant responds. You can test only the current
skill; you cannot test your assistant and all attached skills from end to end. Learn more. |
Dialog |
| Training |
To set up an instance with components that enable the system to function in a particular domain (for example: corpus content, training data that generates machine learning models, programmatic algorithms, annotators, or other ground truth
components) and then making improvements and updates to these components based on accuracy analysis. |
Both |
| Variable |
A variable is data that a customer shares with the assistant, which is collected and saved so it can be referenced later. In actions, you can collect action and session variables. Learn more. |
Both |
| Web chat |
A channel that you can use to embed your assistant in your company website. Learn more. |
Both |
| Webhook |
A mechanism for calling out to an external program during a conversation. For example, your assistant can call an external service to translate a string from English to French and back again in the course of the conversation. Learn more. |
Both |