Summary: ◀▼
You can use AI prompt templates to create customized quality assurance insights that evaluate agent soft skills and customer interactions across industries. These templates help you set criteria for agent behavior and customer risk detection, supporting autoscoring and spotlight insights. You can apply or tailor prompts for categories like communication, compliance, sales, and industry-specific scenarios to improve agent performance and identify customer issues.
Zendesk QA prompt-based categories and spotlights leverage the latest AI models, allowing you to customize AI-powered prompts using natural language for quality autoscoring and risk detection. Rating categories grade agents, while spotlight insights flag tickets.
When creating new AI insights, you can access AI prompt templates by clicking the Prompt templates button.
AI insight prompt templates offer a curated selection of prompts focused on soft skills and organized by industry-specific needs, simplifying implementation. AI Prompt category criteria define agent behavior, while Spotlight criteria are more flexible and can apply to either agent or customer behavior. You can use these templates as they are or customize them to fit your specific requirements.
Related articles
Category prompts list
| Category prompt | Description |
| General | |
| Clear instructions | Did the agent avoid open-ended questions and direct the problem solving with closed questions? |
| Communication style | Did the agent avoid using derogatory words, slang, and complex vocabulary that might confuse the customer? |
| Criticizing the company | Did the agent speak negatively about the company's processes and systems, about the company in general or about some service provider? |
| Question answering | Did the agent answer all the customer's questions in the conversation? |
| Response matching | Did the agent provide a matching response to the customer's questions? |
| Sales insight | Did the agent take the initiative when there were clear signs of a sales opportunity? |
| Solving vs educating | Did the agent guide the customer to a knowledge base, website, or other resources before providing a direct solution to a technical issue? |
| Unnecessary questions | Did the agent request the information that was already available earlier in the conversation? |
| Financial | |
| Customer verification | Did the agent verify the customer's identity before discussing account details? |
| Gaming | |
| Confident communication | Did the agent communicate confidently without using all caps, excessive emojis, grammatical negation, or signs of hesitation such as 'maybe', 'I guess', or 'probably'? |
Spotlight prompts list
| Spotlight prompt | Description |
| General | |
| Competitor mention | Did the customer mention a competitor and claim to have a better offer from somewhere else? Answer yes if all of the above criteria are met. |
| Feature request | Did the customer describe some missing functionality or claim that the current service/functionality is not working properly? Answer yes if any of the above criteria is met. |
| Financial | |
| Signs of fraudulent activity | Did the customer respond with vague or evasive answers, use aggressive language or threats, and request sensitive information when asked specific questions? Answer yes if all of the above criteria are met. Answer no otherwise. |
| Potentially vulnerable customer | Did the customer exhibit signs of vulnerability, such as expressing anxiety, frustration, or distress, and provide inconsistent or unclear responses? Answer yes if any of the above criteria is met. |
| Manufacturing | |
| Quality | Did the customer report a quality issue or express discontent with the product quality? Answer yes if any of the above criteria is met. |
| Retail | |
| Clarification on offers | Did the customer have questions about some offer, such as a promotion, discount, or loyalty program, and ask about how to apply them to their purchase? Answer yes if all of the above criteria is met. |
| Delivery issue | Did the customer claim to have problems with their delivery? |