AiMe Q&A Agent Conversation Logs
You can review all user interactions with AiMe Q&A Agent, which are logged.
AiMe Q&A Agent logs all interactions with users by user and date for audits, quality of service improvements, maintenance, compliance, analytics, and other purposes. Every request and every response is recorded and available for audit. You can export AiMe conversation histories in 10,000-message CSV file batches, filtering by user or date range. You will receive an email notification when an export job is complete. File exports are purged after seven days. The underlying data persists unless deleted.
Logs are retained in comma-separated value (CSV) format and are retained as live data.
Accessing Conversation Logs
Downloading Logs
Use case: As part of an offboarding review, the Compliance team at a pharmaceutical company asks an administrator to produce a record of everything a departing employee asked AiMe over the past year. Opening Advantage Studio from the Conga Start window, the administrator clicks Audit Log, and searches for the departing employee’s user name. The administrator selects the applications the employee used AiMe from and clicks Download. Because the employee was a frequent AiMe user, the resulting Zip file takes several minutes to generate; once complete, it contains the employee’s entire chat history across all selected applications, which the administrator hands off to Compliance for review.
Log Data Analysis
Downloading a log stores a Zip file to your local drive. This file unzips into several comma-separated value (CSV) files, named by each conversation's date and time.
The CSV file format opens in any spreadsheet application, and you can engage with the data using such tools. CSV is also a text-based file format as well, which means that in addition to the data sorting, formatting, and analysis tools available in such applications, you can use text-tool and programmatic data processing tools to analyze the log data you have downloaded.
Data is stored such that Excel and other spreadsheets open each file as a spreadsheet that offers the following basic information about each user conversation.
| Heading | Description |
|---|---|
| Serial Number | This is a running cardinal number series indicating the sequence of messages in the session. |
| Message | This column contains each verbatim question (or other interaction) that the user chatting with the AI entered, and the AI's replies. |
| Time | This is the time of each chat transaction entry from the user, formatted as YYYY-MM-DD 12:59:59.123. |
| Message Type | This field is the record of which party generated the entry: chat_message indicates a user request and assistant_message indicates an AI response. |
| User Feedback | This is a record of the user’s feedback responses. Possible values are:
|
