Release 24.10.3
Date: 11/11/2024
This release includes a bug fix update for SAML secure mail integration
Release 24.10.3
Date: 10/9/2024
Summary
This release includes numerous new features as part of Agent Copilot to assist agents in providing a better client experience, while simultaneously improving productivity and reducing cycle times. The following features are included in this release:
Recommended Responses: Leverage generative AI to provide the best response to an incoming consumer question. These responses are contextually accurate and based on your knowledge base content. Agents can send the recommended response in one click, or edit the response before sending. Agent CoPilot can be configured to either create these recommended responses every time a consumer sends a message or on-demand when agents need help with a recommendation. The key goal is to empower agents with timely and adequate responses to prevent inaccurate responses and speed up the average handle time of the conversation.
Key Conversation Highlights Generate a list of key points mentioned by a consumer during the conversation, relieving agents of the need to take notes during a conversation. These key highlights make it easy for an agent to focus on the conversation, knowing Agent CoPilot will record the notes automatically and accurately.
Knowledge Base AI: Provide a read-only conversational query feature of knowledge base content for agents, to use while conversing with a consumer. This querying feature is available from the agent’s user interface, making it easy to use and get instant answers. Administrators can configure this feature to define one or more knowledge bases, accessible from Agent CoPilot, for each hunt group
Suggested Codes: In situations where LinkLive clients have flat or hierarchical codes, Agent CoPilot will index these codes, and suggest a contextual list of codes matching what has been discussed during a conversation with a consumer, resident, or patient. Agent CoPilot can consume, index, and suggest human services taxonomy codes, medical procedure codes, or any other taxonomy code base. These codes can be first loaded into a LinkLive knowledge base, which will enable Agent CoPilot to retrieve matching codes upon the agent's request.
The key goals of providing these Agent CoPilot features are:
Empowering agents with timely and adequate responses and insights thereby expediting their workflows and reducing their effort leading to higher contact center or customer service team efficiency, as well as an improved agent experience
Preventing incorrect/partial information dissemination by agents, thereby improving consumer experience and retention
Helping onboard new agents faster and retaining existing agents
Please review the following video for a brief walkthrough of these new features.
Note that:
Recommended Responses, Key Conversation Highlights, and Suggested Codes features are only available for chat conversations for now. We will be enabling them for voice calls in the near future.
Recommended Responses, Knowledge Base AI and Suggested Codes need knowledge base(s) first created. This is a separate product which needs to be enabled for your private group and then role based permissions can be assigned. Users with Admin access to the Knowledge Base would be able to create new knowledge base and manage their content. See ‘Knowledge Base Management’ section below for more details.
Updates
Recommended Responses
The recommended response can be either automatically generated for each incoming chat or agents can request a recommended response using the Refresh icon in the UI. The default is on-demand, but this can be configured by supervisors. Agents benefit from on-demand responses when a particular consumer question requires a knowledge base lookup. CoPilot can automatically research and provide a factual, contextual answer and provide a response leveraging the knowledge base retrieved content. For use cases where agents are multi-tasking across multiple chat sessions, an automatic recommended response will be more beneficial.
Once a response is recommended by CoPilot, agents have several choices:
Send the exact recommended response to the consumer as their own reply using the Send Message button.
Edit the recommended response in the text box and then use the Send Message button to send the edited text to the consumer.
Ignore the message and do nothing with it, and simply proceed with their conversation.
Agents also have an option to collapse/expand the recommended response panel if they choose.
A note about Generative AI: LinkLive has gone to great lengths to design Agent CoPilot in a manner that provides governance and guardrails to our AI-enabled features. The recommended responses are contextual and factual based on Knowledge Base content. For non-contextual queries without any content matching in the knowledge base, the recommended response will contain a message informing the agent that your organization does not have a response to the question asked.
Build Number: LinkLive.app 24.09.3
Key Conversation Highlights
Key Conversation Highlights can be either automatically re-generated upon each incoming chat input from the consumer or the agent can request highlights to be generated using the Refresh icon in the UI. Agents benefit from having on-demand highlights generated as conversations build up significantly. This prevents agents from re-reading the chat or taking a mental/physical note of key points. Agent CoPilot will record these highlights automatically, reducing human error, increasing productivity, and improving the agent experience.
Agents have the option to copy these key highlights to their clipboard for use in another system, such as a CRM.
Build Number: LinkLive.app 24.09.3
Knowledge Base AI
The Knowledge Base AI acts as a conversational ChatBot that agents can use to retrieve factual and contextual information related to the questions they receive during chat and voice conversations. This feature provides a quick and easy way for the agent to access the knowledge base(s) relating to the current conversation. Agents can use this interface to perform research while actively working with the consumer during the conversation.
To access the knowledge base AI, agents click into the Knowledge Base tab of the Agent CoPilot and select the knowledge base they want to research. The interface is a simple ChatBot in which the agent can conversationally ask questions.
If the agent needs to switch to a different knowledge base, they can do so by clicking on the selector and selecting a different knowledge base name. Note that when switching between knowledge bases, the questions and answers asked to the previous knowledge base are cleared when switching between knowledge bases.
Build Number: LinkLive.app 24.09.3
Taxonomy Indexing and Code Suggestions
Agent CoPilot can suggest an appropriate taxonomy or code for an agent. This is a great way of retrieving the matching codes to the context consumer described in the chat or call. These codes are retrieved from a knowledge base which is typically built from a CSV file and can help the agent narrow down to the most relevant set of codes matching the user conversation so far.
The suggested codes panel can configured to show only codes or also include descriptions as shown above. The description can be accessed using the expand button. LinkLive will also provide Copy-All and Copy-Selected Code to the clipboard if the agent needs that.
Build Number: LinkLive.app 24.09.3
Knowledge Base Management
One or more knowledge base(s) can be created by uploading your FAQs and other documents on Knowledge Base tab within LinkLive. Each knowledge base can have multiple documents inside it. We support PDF, CSV, TXT, MD and HTML file types and supervisors can drag and drop each file to upload that file into the selected knowledge base. All uploaded documents are processed by LinkLive Knowledge Base AI so that they are ready to provide contextual answers to consumer questions via Agent CoPilot and our Chatbot IVA.
LinkLive offers a conversational knowledge base querying interface, which helps with ad-hoc exploration of the content and is helpful in troubleshooting any potential non-contextual responses. Each query (user prompt) is executed against the selected knowledge base to retrieve the matching content in a ranked order under Query Results.
Build Number: LinkLive.app 24.09.3
Previous Conversations Includes Conversation Summary and Consumer Sentiment
The Previous Conversations widget shows a list of past conversations for the agent. When Consumer Sentiment or Conversation Summary is activated, previous conversations show the Conversation Summary and final Consumer Sentiment score. This allows agents to quickly review historical conversations with authenticated users (chat) and callers (voice), view the final Consumer Sentiment score, and click the conversation to review the detailed transcript.
Build Number: LinkLive.app 24.09.3
Favorite Contacts Widget
The Favorite Contacts Widget enables users to add their most commonly used Contacts into a widget displayed in the Conversations Panel. Once in My Contacts, users can click on a contact to access all capabilities from the contact list including Start Voice, Start Message, and Add to Conference. With My Contacts and the other three widgets, most users will be able to hide the Contact list by clicking the Person icon next to the header search box to open more screen space for messaging and agent tools.
My Favorite Contacts can include:
My Contacts (user added contacts)
Linklive Contacts
Huddles
Conference Bridges
Speed Dials
Agent Connect links
Note that Hunt Group contacts cannot be added to My Favorite Contacts, instead they are added to My Queues. Once added to My Favorite Contacts, users can drag and drop to change the display order.
Users can turn on the Contacts Widget as well as turn on and order all the available widgets in Settings > Communication Center > Widgets. Once activated, users can add Contacts from the contact list by dragging and dropping into the widget or clicking on the contact and selecting ‘Add to My Favorites’.
Build Number: LinkLive.app 24.09.3
Numerous performance and bug fix updates