Conversational Digital Agent
- Chilukuri Srinivasa Reddy (Unlicensed)
- Shilpa K (Deactivated)
- Enterprise IT
- Mayuresh Balaji Kamble (Unlicensed)
What is Conversational Digital Agent?
SummitAI Conversational Interface and Decisioning Engine (CINDE) is a Conversational Digital Agent used to assist the End Users through conversations. The End Users can interact with CINDE to quickly resolve their issues or raise requests in conversational English. CINDE is a digital agent based on the Natural Language Processing (NLP) that understands the user intent, classifies it as Service Request or an Incident, and resolves up to 20% of Service Request tasks in an automated manner.
CINDE supports the JavaScript validations that are configured for Catalogs in the SummitAI application. Using JavaScript support, the fields of the Catalog can be enabled, disabled, or pre-filled. Based on the data provided in the text fields, the validation messages pop-up in the SummitAI application. The same concept is applied in CINDE in a conversational way.
When a user asks CINDE for a Catalog that is configured with JavaScript, the fields configured in the Catalog are displayed as questions. Based on the provided information, the JavaScripts are executed and CINDE responds accordingly with subsequent questions. The pop-up validation messages displayed in the SummitAI application are displayed as instructions in CINDE for the user. For more information about JavaScript Validations, see Creating Service Catalogs in Service Management.
CINDE is available as multiple-channel support that can be invoked by clicking the icon. On clicking the icon, a chat window is displayed, where the End Users can type in their questions or the intent. CINDE responds to the questions or the intent related to Incident Management and Service Request Management.
Users can click the icon on the top-right corner of the CINDE window to provide feedback about their conversation with CINDE. The user is redirected to the configured mail box and after a feedback mail is sent, it is parsed into an Incident.
Live Agent Transfer
CINDE can transfer conversations to an Analyst in instances where either CINDE is unable to provide correct solution to the End User's queries or if the End User wants the conversation to be transferred to an Analyst. This improves customer satisfaction where the user queries are transferred directly with all the conversation history to a live Analyst, which helps in providing quick solutions to the users. Users can click icon on the top-right corner of the CINDE interface or ask CINDE to transfer the conversation to a Live Agent. For more information, see Live Agent Transfer.
Features of CINDE
Incident Management
Using CINDE, End Users can do the following:- Log Incidents
- Update Incidents with more information
- Check Incident Status
- Remind Analysts to resolve Incidents
- Cancel Incidents
- Escalate Incidents
Click to see examplesService Request Management
Using CINDE, End Users can do the following:- Log Service Requests
- Update Service Requests with more information
- Check Service Request Status
- Remind Analysts to resolve Service Requests
- Cancel Service Requests
- Escalate Service Requests
Click to see examples- Integrating CINDE with Incident Rules
Automation and intelligently identifying user's requirement is the key to the success of Artificial Intelligence (AI). Taking a step in this direction, CINDE's integration with Incident Rules will help the end-users to log incidents very quickly without spending much time in describing the problem statement. This will also help Analysts to get the Incident assigned to the appropriate Workgroup with all the pre-filled Incident details such as Priority, Urgency for a faster closure of the incident. It provides capability to display the default symptoms based on the keyword specified by the end-user.
Click here for detailed information and examples
Intelligent Recommendations
CINDE provides intelligent recommendations in the form of Knowledge Articles or multimedia content to resolve the issue before logging any Incident/Service Request. These intelligent recommendations also reduce the number of Incidents/Service Requests logged by the End Users. If KB articles cannot resolve the issue, the user can continue to create an Incident/Service Request to log the issue.Click to see examplesChit Chat
The End User can chit chat with CINDE using conversational English. Based on the user's input, CINDE answers and provides options that the user can choose from (For Example: Create a ticket or Call Helpdesk).Click to see examplesFrequently Asked Questions
CINDE answers the Frequently Asked Questions asked by End Users. This helps the End Users to get prompted with the most probable issues, which they can select and also find a solution before logging any Incident. This helps in quick and easy resolution of issues and also reduces the number Incidents logged by the End Users. With this, the users can ask CINDE for information and get answers in a conversational way. The Frequently Asked Questions (FAQs) output is specific to the respective organization based on the data input provided to CINDE.
Using this feature, the user can quickly find answers to the common IT and Services related questions. To train and make the system proficient, the organizations must provide a set of organization-wide FAQs (and variants of these questions) as well as their answer pairs.Note:
These FAQs are updated periodically based on the inputs from the IT Admin to SummitAI team. Training/re-training CINDE can take few days.
Click to see examples
Advantages of CINDE
- CINDE is available 24*7 – End Users can interact with CINDE anytime, anywhere.
- Users no longer need to navigate through multiple pages of the portal to look for Knowledge Records or to log Incidents or Service Requests. CINDE helps with all of these with a click of a button.
- CINDE enhances the user experience with significant reduction in resolution time.
- CINDE reduces back and forth conversations between user and analyst.
- CINDE improves Self Service with its intelligent recommendations that can be leveraged by End Users to resolve the problems themselves.
- Omni-channel Support offers a convenient way for users to get the work done quickly.
- Improves productivity: CINDE enables Analysts to focus on more complex tasks by answering and resolving common requests made by users.
- CINDE reduces costs by handling higher volumes of repetitive tasks.
Multi-Channel Support
CINDE provides 24*7 support to the End Users and provides quick solutions to the End Users. Along with the SummitAI application, CINDE is available as a contact for the configured users on the following applications:
- Web chat
- Microsoft Teams
- Slack
- Jabber
With this, the users can perform the actions related to Incidents or Service Requests directly from any of these channels without logging into the SummitAI application.
Spell Check
CINDE checks the spelling of user utterances (typo errors, etc.) to provide more useful responses to user interactions. The user can get suggestions for the entered utterance if there are any misspelled words. When both the suggestion and intent are same, the user can select CINDE suggestion by clicking Yes and continue with the conversation. If the suggestion and intent are not same, the user can click No, use original, and proceed with the entered utterance. See the following screenshot as an example:
Location Based User Access
The user can access CINDE based on the configured Location in the Summit_Config table with the comma separation. For example, see the below reference:
IF NOT EXISTS(select 1 from Summit_Config where Config='Summit_Assistant_EnabledLocations' and Active=1) BEGIN INSERT INTO Summit_Config (Org_ID,Config,Config_Val,Active) values(1,'Summit_Assistant_EnabledLocations','Location 1,Location 2',1) END ELSE BEGIN UPDATE Summit_Config SET Config_Val = 'Location 1,Location 2' where Config='Summit_Assistant_EnabledLocations' and Active=1 END GO
Note:
The Location IDs mentioned in the above example are for demonstration purpose.
Confluence Cloud Migration Alert: Please refer to known issues you may encounter in Confluence Cloud: https://eitdocs.atlassian.net/wiki/x/wDGwAQ