# genai_extract_user_prompt



The `genai_extract_user_prompt` function extracts the user’s prompt from a GenAI messages array. It returns the content of the last message with the 'user' role, which typically contains the user’s question or request to the AI.

You can use this function to analyze user queries, understand common question patterns, perform sentiment analysis on user inputs, or track user behavior and needs.

## Usage [#usage]

### Syntax [#syntax]

```kusto
genai_extract_user_prompt(messages)
```

### Parameters [#parameters]

| Name     | Type    | Required | Description                                                                                                         |
| -------- | ------- | -------- | ------------------------------------------------------------------------------------------------------------------- |
| messages | dynamic | Yes      | An array of message objects from a GenAI conversation. Each message typically contains `role` and `content` fields. |

### Returns [#returns]

Returns a string containing the content of the last user message in the conversation, or an empty string if no user message is found.

## Example [#example]

Extract the user's prompt from a GenAI conversation to analyze common questions.

**Query**

```kusto
['otel-demo-genai']
| extend user_query = genai_extract_user_prompt(['attributes.gen_ai.input.messages'])
| where isnotempty(user_query)
| summarize query_count = count() by user_query
| top 5 by query_count
```

[Run in Playground](https://play.axiom.co/axiom-play-qf1k/query?initForm=%7B%22apl%22%3A%22%5B%27otel-demo-genai%27%5D%20%7C%20extend%20user_query%20%3D%20genai_extract_user_prompt\(%5B%27attributes.gen_ai.input.messages%27%5D\)%20%7C%20where%20isnotempty\(user_query\)%20%7C%20summarize%20query_count%20%3D%20count\(\)%20by%20user_query%20%7C%20top%205%20by%20query_count%22%7D)

{/* vale off */}

**Output**

| user\_query                   | query\_count |
| ----------------------------- | ------------ |
| How do I reset my password?   | 456          |
| What are your business hours? | 342          |
| How can I track my order?     | 298          |

{/* vale on */}

This query identifies the most common user questions, helping you understand user needs and improve responses.

## List of related functions [#list-of-related-functions]

* [genai\_extract\_assistant\_response](/apl/scalar-functions/genai-functions/genai-extract-assistant-response): Extracts the assistant's response. Use this to analyze AI responses along with user prompts.
* [genai\_extract\_system\_prompt](/apl/scalar-functions/genai-functions/genai-extract-system-prompt): Extracts the system prompt. Use this to understand the AI's configuration when analyzing user queries.
* [genai\_get\_content\_by\_role](/apl/scalar-functions/genai-functions/genai-get-content-by-role): Gets content by any role. Use this for more flexible extraction when you need other specific roles.
* [genai\_concat\_contents](/apl/scalar-functions/genai-functions/genai-concat-contents): Concatenates all messages. Use this when you need the full conversation instead of just the user prompt.
* [genai\_estimate\_tokens](/apl/scalar-functions/genai-functions/genai-estimate-tokens): Estimates token count. Combine with user prompt extraction to analyze prompt sizes.

## Other query languages [#other-query-languages]

<LanguageComparisons>
  <Accordion title="Splunk SPL users">
    In Splunk SPL, you would need to filter messages by user role and extract the last one.

    <CodeGroup>
      ```sql Splunk example
      | eval user_msgs=mvfilter(match(role, "user"))
      | eval user_prompt=mvindex(user_msgs, -1)
      ```

      ```kusto APL equivalent
      ['ai-logs']
      | extend user_prompt = genai_extract_user_prompt(messages)
      ```
    </CodeGroup>
  </Accordion>

  <Accordion title="ANSI SQL users">
    In ANSI SQL, you would unnest the array, filter by user role, and select the last message.

    <CodeGroup>
      ```sql SQL example
      SELECT
        conversation_id,
        content as user_prompt
      FROM (
        SELECT *, ROW_NUMBER() OVER (PARTITION BY conversation_id ORDER BY msg_index DESC) as rn
        FROM conversations
        CROSS JOIN UNNEST(messages) WITH OFFSET AS msg_index
        WHERE role = 'user'
      ) WHERE rn = 1
      ```

      ```kusto APL equivalent
      ['ai-logs']
      | extend user_prompt = genai_extract_user_prompt(messages)
      ```
    </CodeGroup>
  </Accordion>
</LanguageComparisons>