Agent Assist Use Case Map

#

Customer Use Case

Agent Assist Feature

Input Data

Data Source Utilized to Generate Answer

Key KPIs

1

Save agents time from summarizing the issue, actions taken and resolution for each case.

Case Summary - Succinct summary of the primary issue reported, case progression and resolution provided available in the Aisera Assist widget.

Current case Title, Description, Comments and the specified fields

N/A

Agent Copied

Case Summary Feedback

2

Helping the agent draft a case response.

Aisera Answer - Generated response to the user's query based on the knowledge content Aisera has access to. Available in the Aisera Assist widget.

Current case Title, Description, Comments and specified fields

Customer Knowledge Base Articles & Knowledge Articles generated from historical cases

Agent Applied

Agent Clicked

Aisera Answer Feedback

3

Aisera provides recommendations on next steps for troubleshooting cases.

Next Best Actions - Troubleshooting step recommendations available in the Aisera Assist widget.

Current case Title, Description, Comments and specified fields

Historical Cases

Agent Copied

Next Best Action Feedback

4

Help agents find relevant information to resolve customer issues.

AI Assistant - Conversational chatbot available for Support Agents to ask questions on ingested knowledge articles in the Agent Assist widget.

User query in chatbot

Customer Knowledge Base Articles & Knowledge Articles generated from historical cases

Resolution Rate

# of Requests

Adoption Rate

5

Reduce manual effort required by Agents to categorize cases

Field Predictions -

Aisera’s AI predicts values for specified ticket fields available for an agents’ review in the Agent Assist widget

Case Title and Description

N/A (Learn from historical tickets)

Prediction Accuracy

The Case Summary generates a condensed version of the current case by highlighting the main issue(s) reported, actions taken and resolution provided. The case summary will utilize the case Title, Description, Comments and specified fields to generate the result. The summary will evolve as the case content evolves. The case summary provides agents with a comprehensive overview, allowing them to grasp the case's intricacies and resolution without needing to sift through extensive comments.

You have the ability to

1. Copy the case summary to their clipboard for future use, such as, providing status updates on an escalated ticket or creating content.

2. Provide Feedback - It is crucial for agents to provide feedback on the Case Summaries generated by Aisera. This information will be captured in the Analytics Dashboard and be utilized to measure the accuracy of Aisera’s recommendations and to improve Aisera’s recommendations.

   
3. Upon selecting Thumbs Down, users will have the option to select the reason the recommendation was not useful and provide additional feedback in an open form textbox. The labels for the dropdown and textbox and the feedback dropdown options are customizable.

Case Summary KPIs

Available in the Aisera Admin UI - Agent Assist > Ticket Agent Assist Tab

Agent Copied - The number of times an agent copied the recommended Case Summary.

Case Summary Feedback - The amount of Thumbs Up and Thumbs Down provided on Case Summary.

Negative Feedback Reason - The reason the agent entered for why the recommended summary was not helpful.

Aisera Answer generates a natural language response to the current case based on the knowledge content ingested into Aisera. The answer aims to serve as a direct response that Agents can send to the case reporter. Answer generation saves agents time from searching for the answer, summarizing it and sending it to the customer.

You have the ability to

1. Copy the answer and paste it as a draft comment. You can edit the response to tailor it to their liking or directly post the answer as a comment.

2. Provide Feedback - It is crucial to provide feedback on the Answers generated by Aisera. This information will be captured in the Analytics Dashboard and be utilized to measure the accuracy of Aisera’s recommendations and to improve Aisera’s recommendations.

   
3. Upon selecting Thumbs Down, users have the option to select the reason the recommendation was not useful and provide additional feedback in an open form textbox- we suggest asking agents to provide the link to the knowledge article or historical case that they utilized to solve the case. The labels for the dropdown and textbox and the feedback dropdown options are customizable.

Aisera Answer KPIs

Available in the Aisera Admin UI - Agent Assist > Ticket Agent Assist Tab

Agent Applied - The number of times an agent applied Aisera Answer to the ticket.

Agent Clicked - The number of times an agent clicked into the reference articles provided by Aisera Answer.

Aisera Answer Feedback - The amount of Thumbs Up and Thumbs Down provided on Aisera Answers.

Negative Feedback Reason - The reason the agent entered for why the recommended answer was not helpful.

Next Best Actions generates troubleshooting steps based on Agent Resolution Notes from historical tickets. The Next Best Actions feature suggests actions that were successful in similar cases. A Next Best Action card consists of an Issue, Root Cause and Troubleshooting steps from a historical case. Multiple Next Best Action recommendations may be relevant to a single ticket. In this case, Aisera will recommend various troubleshooting steps based on slightly different issues and root causes from similar cases.

You have the ability to

1. Copy the content so you can edit the response or directly post the comment for the reporter.

2. Previous/Next - If multiple sets of root causes and troubleshooting steps are recommended, you will have the option to navigate back and forth between the recommendations.

3. Provide Feedback - It is crucial for you to provide feedback on the Next Best Actions recommended by Aisera Assist. This information will be captured in the Analytics Dashboard and be utilized to measure the accuracy of Aisera’s recommendations and to improve Aisera’s recommendations.

   
4. Upon selecting Thumbs Down, you have the option to select the reason the recommendation was not useful and provide additional feedback in an open form textbox. The labels for the dropdown and textbox and the feedback dropdown options are customizable.

Next Best Actions KPIs

Available in the Aisera Admin UI - Agent Assist > Ticket Agent Assist Tab

Agent Copied - The number of times an agent copied the Next Best Actions to the ticket.

Next Best Action Feedback - The amount of Thumbs Up and Thumbs Down provided on Next Best Actions.

Negative Feedback Reason - The reason the agent entered for why the recommended next best action was not helpful.

Aisera Assistant is a conversational chatbot available to help Customer Support Agents solve complex cases without leaving their System of Record.

- Search across various knowledge sources with a single request and receive a natural language summarized answer to their question. (For example: Salesforce, SNOW Knowledge Base)

- Ask questions that can be answered by public knowledge articles that are whitelisted in your Aisera App configuration. (For example: How do you connect your gmail to Slack? What is the latest upgrade available for Mac?)

- Ask the assistant for writing and editing assistance, such as-

• Draft an email response to the customer/employee using the recommended summary and answer

• Make an aisera answer or the summary more succinct, professional, rewritten in bullet points, etc.

• Remove certain sections of an answer or summary

• Ask for a case wrap up

- Execute action workflows from the assistant (For example: Provision software access)

Suggested Queries/Best Practices:

Hey Assistant, draft me an email response to an escalated customer about a token expiry issue.

1. Hey Assistant, can you make the answer more succinct and use bullet points

2. Once you receive the bulleted answer, you can engage back asking for additional changes such as → Rewrite this in a more friendly tone.

3. When asking questions that can be answered by the knowledge base: a) Do not use 1 - 2 word queries b) Keep it relatively concise c) Provide enough context for the assistant to understand what you are looking for

1. Provide feedback on the responses you receive!

Aisera Assistant KPIs

Available in the Analytics > Pre-Built > Requests Tab and Users Tab

Resolution Rate - The number of resolved and assisted requests divided by the number of relevant requests (Goal: 80%+)

Total Requests - The number of requests entered in the assistant

Adoption Rate - The number of total active users divided by the potential number of users

Aisera provides Field Prediction recommendations through the Aisera Assist widget, designed to help agents resolve cases more efficiently and increase the accuracy of case categorization. By analyzing the ticket title, description, and specified fields, Aisera predicts the values for desired fields, including a confidence score to indicate the certainty of each recommendation. This prediction is generated when the Aisera Assist widget loads. The prediction updates as additional information is added to the case.

Agents can apply the predicted value directly to a field by selecting the "Apply" checkbox. Upon selection, a success message will appear on the screen, and the predicted value will populate the field. If the field already contains a value, it will be replaced when "Apply" is pressed. Currently, deselecting the checkbox does not remove the prediction or revert the field to its previous value.

Best Practices-

1. Aisera improves its predictions by learning from historical case data. Therefore, maintaining a clean and reliable dataset is essential. Predictions for values with low data volume will have lower confidence scores.

2. If you plan to make any Taxonomy changes, it is imperative to notify Aisera ahead of time so we can retrain our models based on the new field values.

Field Prediction KPIs

Available in the Aisera Admin UI - Agent Assist > Predict & Audit Tab

Main KPIs-

Prediction Accuracy - the number of correct predictions / the number of auto-applied predictions (Goal: 80%+)

Correct predictions are determined by comparing the field value Aisera predicts with the Actual field value on the case, if this value is changed, it will be reflected in Aisera when the next ticket ingestion happens, typically nightly.

Coverage Rate - the number of auto-applied tickets / the number of predicted tickets (Goal: 80%+)

Supporting Metrics-

Distinct Values - The number of distinct values that Aisera has recommended for the specified field

Predicted Tickets - The total number of tickets that Aisera made predictions for

Auto Applied Tickets - The number of tickets where Aisera's field prediction met the confidence threshold defined in the configuration tab

Correct Predictions - The number of tickets where Aisera's field prediction value equals the actual field value on the ticket