Bi-Weekly Product Release Notes - Sept 6, 2024
New Features
The following features were added to the Aisera platform in this release.
Agent Assist
The following features have been added to Agent Assist in this release.
Added Case Summary and Next Best Actions to the Ticket Details Page
You can now click into an individual ticket to view the Case Summary, Aisera Answer and Next Best Actions that were recommended for the ticket. As recommendations evolve, the latest suggestion are displayed.
Added the Negative Feedback Reasons to the Agent Assist Analytics Page
When an agent presses thumbs down on a Recommendation in Agent Assist, they have the option to select the reason that the recommendation was not helpful. This Feedback is visualized in a new pie chart, accessible in the Aisera Answers, Next Best Actions and Case Summary sections of the Agent Assist Analytics page.
Updated the Analyzed Tickets KPI Calculation and Definition
The calculation and definition for the Analyzed Tickets KPI has been refined.
As of this release, it represents the number of distinct tickets where an agent opened the case in the ticketing system and then the Aisera Agent Assist widget loaded.
Updated the Predicted Tickets KPI Calculation and Definition
The calculation and definition for the Predicted Tickets KPI has been refined.
Now, it represents the number of distinct tickets where Aisera made has at least one non-empty prediction for Case Summary, Aisera Answer, Next Best Action, or Field predictions.
Updated the Tickets with Agent Action KPI Calculation and Definition
The tickets with Agent Action KPI calculation and definition now reflect the number of distinct tickets where an agent provided Feedback, copied the Summary, Answer or Next Best Action or applied a Field prediction. Expanding or collapsing recommendation sections no longer impacts this metric.
Streamlined the Agent Assist Analytics Page
We’ve decluttered the Agent Assist analytics page by removing retired charts.
We updated labels to reflect the product name as Agent Assist. Additionally, we’ve reorganized the page to prioritize analytics based on usage patterns.
Agent Assist in Zendesk
We are excited to announce that Aisera's Agent Assist is now available in Zendesk. This integration brings powerful AI-driven capabilities directly to the Ticket Details page, enabling agents to resolve cases more efficiently.
With Agent Assist embedded in Zendesk, agents can now benefit from:
Field Predictions: Accurate recommendations for predicting key fields, saving time and reducing errors.
Ticket Summarization: Instant summaries of tickets to quickly understand the context and details
Answer and Action Recommendations: Relevant suggestions drawn from Knowledge Articles and historical cases, streamlining the resolution process.
Conversational Chatbot: An AI-powered chatbot that assists agents with queries and provides instant support.
For more information, please reach out to your Customer Success Manager.
Conversation AI 1.0
The following features have been added to Conversation AI 1.0 in this release.
Workflow as NLU Pipeline
You can now define and select a Workflow as a Fulfillment pipeline and add it as a segment anywhere within the existing pipeline. The Workflow can fulfill the user’s Request or pass the question on to the next pipeline.
To Set Up Workflow as an NLU Pipeline:
Use the Workflow Studio to create a Workflow to serve the Request, with the parameters shown in the diagram below.
2. Open the Advanced settings window for a new or existing application/bot.
3. Select Workflow in the Fulfillment Engines field.
4. In the Flow field, select the flow you want to use from the pull-down list.
Enhancements
The following enhancements have been added to the Aisera platform in this release.
Hyperflows
The following enhancements have been made to Hyperflows in this release.
Hyperflows Include Workflow Functions
The Hyperflow function can now be either a Workflow or an API.
Workflow is now a third type of Function for Hyperflows.
Hyperflow Authentication for APIs
Hyperflows include the ability for flows to execute the APIs/Actions associated with an integration as the user who is logged into the channel of interaction. For instance, if you are using a bot in the Aisera Webchat channel, the API or Action authentication will be performed for the user logged into that channel.
New Hyperflow Status
If the Hyperflow Engine can not continue a conversation based on what the user has provided, it returns a special out of scope status. After returning an out of scope status, the conversation server goes back to determine if there is any other fulfillment that it can serve to the user.
Workflow Studio
The following enhancements have been made to Workflow Studio in this release.
New LLM Image Action for Workflow Studio
You can now add an image that is valid input for GPT API questions.
In previous releases, you could choose LLM Chat Answer as an Action Node for use with the LLM Gateway. Now this option has been improved as ‘LLM Chat Answer 2’ with an easier setup configuration and the ability to take images as input. Input options for images include public URLs and Base64 encoded images.
To Use the LLM Chat Answer 2 Option:
Select AI Automation > AI Workflow Studio.
Select AI Automation > AI Workflow Studio.
Choose +New Workflow or open an existing Workflow.
Drag an Action node onto the Workflow Studio canvas.
Choose Action Categories from the View Actions By field.
5. Specify your input parameters and variables in the Edit Action Node window.
Set Message Node for Copilot Markdown Variable
You can now use a Workflow Message Node to designate that the value of a variable is in Markdown format.
To designate that the value of a variable for a Copilot application is in Markdown format:
Create a new application/bot or open an existing application/bot in the Aisera Admin UI.
Select AI Automation > AI Workflow Studio.
Drag a Message node onto the Workflow Studio canvas.
Specify that the value of the variable is in Markdown format, as shown below:
Conversational AI 2.0
The following enhancement has been made to Conversational AI 2.0 in this release.
Exit Action Flow and Resume
With this release, you can now exit an Action flow by changing the context (such as sending a new Request) and then return to the flow exactly where you left off.
Example Use Case: You enter an enrollment flow and are provided with a form to fill in. When you send a second request (such as, “What does this field mean?”) you immediately exit the flow, get served for the new Request (provided that there is relevant information), and then resume to finish your enrollment.
To enable this feature, turn on the Show UnFinished Context Hint option in the Settings > Configuration > ContextManagement window.
NOTE: This feature is not available for steps in a Workflow where the user is asked to type text.
Last updated