Objective
To automate the initial evaluation of inbound support cases by utilizing AI-powered sentiment analysis, aiming to improve response times for critical issues and overall customer satisfaction.
Pre-conditions
- The native app, Q-assign, is installed in Salesforce (or, alternatively, Email-to-anything*)
- Ortoo’s Plug-in Actions feature is enabled in Q-assign.
- Customer support agents must be trained on handling cases post-sentiment analysis.
- Inbound cases must be properly logged and categorized within Salesforce.
Actors
- Salesforce Administrator
- AI Sentiment Analysis Tool
- Customer Support Agents
- Inbound Case (as an entity)
Workflow
- An inbound case is created in Salesforce when a customer submits a support request.
- Q-assign’s Plug-in Action is triggered upon the case’s arrival, initiating the sentiment analysis.
- The AI tool processes the case content, assessing for positive, neutral, or negative sentiment.
- The case is then automatically tagged with the sentiment score and categorized accordingly within Salesforce.
- Based on the sentiment analysis, cases with negative sentiment are prioritized in the queue.
- Customer support agents are alerted to new, high-priority cases requiring immediate attention.
- Agents respond to prioritized cases, equipped with the insight provided by the sentiment analysis.
Post-conditions
- Cases are sorted and prioritized based on their sentiment score.
- Support agents are immediately aware of high-priority cases.
- Customers receive timely responses, especially for urgent or negative experiences.
Exceptions
- If the sentiment analysis tool is unavailable, the case is flagged for manual review.
- Cases with mixed sentiment signals are flagged for secondary review to determine priority.
Metrics for Success
- Decrease in average response time for high-priority cases.
- Increase in customer satisfaction scores (CSAT).
- Reduction in the number of escalations due to delayed response.
- Improved efficiency in case handling and resource allocation.
Enhancements: Further developments could include refining AI models for greater accuracy in sentiment detection and expanding the analysis to predict potential resolution strategies.
