Skip to content

Commit bba0913

Browse files
Generate NLU Utterances from Intent.md (#52)
* Generate NLU Utterances from Intent.md This prompt addresses the challenge of manually creating varied Utterances for ServiceNow NLU Intents. Additionally: - It provides a structured approach for generating 15 distinct user utterances. - It includes instructions, steps, and examples to ensure coverage of different communication styles and common user mistakes. * Generate NLU Utterances from Intent.md This prompt addresses the challenge of manually creating varied Utterances for ServiceNow NLU Intents. Additionally: - It provides a structured approach for generating 15 distinct user utterances. - It includes instructions, steps, and examples to ensure coverage of different communication styles and common user mistakes.
1 parent 0b45a88 commit bba0913

1 file changed

Lines changed: 33 additions & 0 deletions

File tree

Lines changed: 33 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,33 @@
1+
### Role:
2+
You are a ServiceNow developer tasked with improving the Natural Language Understanding (NLU) and Virtual Agent functionality.
3+
Your goal is to generate diverse and realistic user utterances for specific intents to enhance training data for the AI models.
4+
5+
### Instructions:
6+
You are to create 15 unique utterances (sample user statements) for the intent provided below.
7+
These utterances should represent a variety of ways users might express the same request or problem when interacting with ServiceNow.
8+
Be sure to account for different communication styles, user expertise levels, and potential typos or mistakes that real users may make.
9+
10+
### Intent:
11+
**[YOUR_INTENT_HERE]**
12+
13+
### Steps:
14+
1. Identify the core action or request related to the intent.
15+
2. Create 15 different utterances reflecting how users might phrase the request, using both formal and informal language styles.
16+
3. Incorporate both short and longer utterances, while reflecting different user levels (beginner to expert).
17+
4. Ensure that the utterances account for possible typos or slightly incorrect phrasing, as real users often make mistakes.
18+
19+
### Example:
20+
21+
For the intent **"Reset Password,"** some sample utterances could include:
22+
23+
1. "I forgot my password, can you help me reset it?"
24+
2. "How do I reset my ServiceNow password?"
25+
3. "Password reset request."
26+
4. "Can you help me recover my password?"
27+
5. "I need to reset my password because I forgot it."
28+
29+
### End Goal:
30+
Provide a diverse set of sample user utterances to enhance the training of a ServiceNow NLU model or to create a more responsive Virtual Agent experience.
31+
32+
### Narrowing:
33+
Limit each utterance to **40 words or fewer**. Include at least 3 examples that represent common user mistakes in spelling or grammar.

0 commit comments

Comments
 (0)