Bot logs are available for 30 minutes after the last message in the conversation. This is the case both during preview in bot development and in a conversation on any channel. This TTL for bot logs isn't related to the bot's session length, and the value isn't configurable.

Introduction

As you design and implement a bot, you can use the Preview and Bot Logs tools, respectively, to preview the conversational experience and to debug issues. These tools assist you with early testing and debugging, so you can verify the conversational flow is working as expected as you build the bot. Notably, the tools don’t require that you deploy an agent connector for the bot because the conversation doesn’t go through Conversational Cloud. The conversation only flows between the tool and the underlying bot server.

In contrast, the Conversation Tester tool is designed for testing and debugging after deployment. You can use the tool to:

• Test the end-to-end flow after you have finished building and deploying a bot
• Debug a deployed bot

The Conversation Tester lets you quickly start a conversation and view both the conversation and the debug log side by side.

With the Conversation Tester, the conversation travels from the Conversational Cloud Web channel, to Conversational Cloud, to the agent connector, to the bot server, and back again. In short, there is end-to-end testing of the entire flow over the Web.

Currently, you can’t use the Conversation Tester to test post-conversation survey bots.

Prerequisites

To use the Conversation Tester, the entire pipeline for messaging must be in place and ready:

• In Conversational Cloud, there must exist a bot user (an agent) that is assigned a skill.
• In LivePerson Conversation Builder, the bot must have a deployed agent connector that's assigned to the bot user, and the connector must be online.

Access the Conversation Tester

You can access the Conversation Tester in two ways:

• From within a custom bot, click on the menu bar, and select Conversation Tester.

  This option is available only when the bot has an agent connector (regardless of its status). When you want to test a specific bot, this method is quicker because it automatically starts a conversation with that bot, routing it to the first, associated skill that is found.

• From the bot dashboard, click Conversation Tester on the menu bar in the upper-left corner.

  When you want to perform end-to-end testing (of the engagement, skill, bot agent, and bot), this method is recommended because you'll be starting the conversation by specifying the engagement to use, not the bot to use.

  Access via this method does also allow you to select a skill by bot, or by skill directly.

Start the conversation

Conversational Cloud routes conversations based on skill, so to start a conversation, you need to specify the skill to which to route it.

If you access the Conversation Tester from within a bot, the tool automatically starts a conversation with that bot, routing it to the first, associated skill that is found.

If you access the Conversation Tester from outside of a bot, you'll need to explicitly select the skill and manually start the conversation.

To select the skill and start the conversation

1. Click New Conversation in the upper-right corner.

   

2. Select the skill. You do this in one of three ways:

   

   • Select skill: Select the skill. The skill list displays all the skills available within your organization. If you know the skill name to route to, this one-step method is fastest.

   • Select skill by bot: Select the custom bot, then the skill. For a bot to appear in the bot list, it must be deployed and online. Once you select the bot, the skill list is filtered to display only the skills assigned to the associated bot agent user. This two-step method is preferable when you want to route the conversation to a specific bot, but you aren't sure of the skill name.

   • Select skill by campaign/engagement: Select the campaign to use. This filters the list of engagements from which you must then select the engagement. The skill tied to the engagement will be used to start the conversation. This method is preferable when you want to perform end-to-end testing of the entire Conversational Cloud configuration: engagement, skill, bot agent, and bot.

3. Click Start Conversation.

Start the debugger

Click the debug icon in the lower-left corner of the messaging panel.

This starts the debugger:

Use the messaging panel

Send messages

Once the conversation is started, you can send messages in the messaging panel to direct the conversational flow. This works much like Preview.

Close the conversation

To close the current conversation, click the icon at the bottom of the messaging panel.

Reset the bot

To reset the bot, click the icon at the bottom of the messaging panel. This deletes the bot session, so you can start the conversation flow anew.

Use the debug panel

Log debug messages

Keep the Auto update setting on if you want to continuously fetch the latest logs based on the conversation.

If you don't need the log, you can turn Auto update off, which keeps the connection clean. In this case, you can manually refresh the log by clicking .

Show and hide debug messages

The debug log can become verbose and long, so it can be helpful to show and hide specific types of messages. Click and use the available filters to do this.

Search the debug log

To search the debug log for a partial word, word, or phrase, enter it in the search box that’s provided.

Copy the user ID

In scenarios where you’re troubleshooting an issue in collaboration with LivePerson Support, you might be asked to provide the user ID for the conversation. There are two ways you can obtain the ID:

• Click the Copy user ID button that’s displayed in the messaging window when you start the debugger.
• Click the Copy user ID icon at the bottom of the debugging panel. This too is displayed once you start the debugger.

Print the most recent user message

To aid in debugging, you can use the printDebugMessage scripting function in the code areas of an interaction to print messages — for example, the consumer's most recent message — to the debug log.

Events glossary

To assist you as you debug, refer to this events glossary.