1. Introduction to the AI Configurations Page
Welcome to the AI configurations page. This area is divided into four main sections: Live, Config, Memories, and Automation.
2. Exploring the Live Tab
Let’s explore each section. In the Live tab, you can view all logs of the AI's responses. On the left side, you can either search by a specific guest name or use the dropdown to choose a particular listing.
3. Filtering and Analyzing Logs
You can filter logs by hour, 12 hours, day, or week. By clicking on More Filters, additional options like conversation ID, log type, AI generation type (Autopilot or Copilot), agent type, and specific date filters are available. This area is particularly useful for troubleshooting and understanding AI responses.
4. Overview of the Config Tab and AI Agents
Now, let’s move to the Config tab. Here, you will find five AI agent types: reviews, customer support, sales, operations, and owner relations.
Differentiating AI Agents
A. Sales Agent
Sales agents focus on potential guests whose inquiries are still in the Inquiry stage. The primary role is to guide these prospects, answer questions about availability or pricing, and convert inquiries into confirmed reservations.
B. Customer Support Agent
The Customer Support agent manages interactions related to confirmed bookings. They address guest questions about ongoing stays, resolve issues during visits, and provide post-stay support. Their goal is to ensure guest concerns are handled quickly and effectively.
C. Operations Agent
Operations agents interact primarily with contractors and property managers. They are responsible for creating service calls and assigning maintenance tasks. Unlike Customer Support agents, they do not handle questions from current guests.
D. Owner Agent
The Owner agent manages inquiries related to property performance and reporting.
5. Customer Support Agent Configuration Overview
These agents have similar configuration screens, so we will focus on Customer Support, the most commonly used. Within this agent, there are several sections such as General, Automation, Communication, Security, Policies, and Advanced. In the General tab, the AI Copilot can be enabled.
6. Enabling AI Copilot and Debug Options
Enabling the AI Copilot allows it to draft responses that require human approval before being sent. Below this setting, you can enable Save Chain of Thought to make AI reasoning visible in the Live tab and Debug Mode for additional logs.
7. Additional General Settings
You can deny store links to prevent the AI from including store item links in responses. The option to customize the agent name is also available here.
8. Log Rejection and Tagging Features
The Log Reject Reason feature requires users to provide a reason when rejecting an AI-suggested response, which appears as a prompt in the CRM. Tag Conversation enables the AI to automatically tag conversations.
9. Conversation Tagging and Formatting Settings
For example, when a guest shows satisfaction, these tags become visible in the Memories tab. Settings for country, date format, time format, and distance format control how AI responses are formatted.
10. Automation Section Overview
Next, let’s explore the Automation section. Here, you can activate Autopilot, allowing the AI to send responses automatically without human intervention.
11. Follow-Ups and Escalation Controls
You can configure the AI to follow up with guests to confirm response helpfulness and set the number of follow-ups. The Pause on Escalation feature stops Autopilot if a conversation is escalated, preventing further autonomous replies.
12. Autopilot Scope and Shift Scheduling
You can specify which listings have Autopilot enabled and set a Shift Schedule to limit automatic AI responses to specific hours. In the Communication tab, the AI’s response style can be configured.
13. Communication Style Customization
Adjustments can be made to response length (concise, detailed, balanced), formality (casual, formal, balanced), and empathy level (direct, empathetic, balanced). A signature can also be added and toggled as needed.
14. Security and Policies Overview
In the Security tab, two key settings ensure the pre-check-in form is completed and the deposit is paid before the AI provides door codes. The Policies tab is crucial for defining response rules.
15. Creating New Policies
The AI uses policies to ensure responses align with your standards. To add a new policy, click Add New Topic. A modal will appear where you can enter the topic and subject name.
16. Writing and Managing Policy Content
Once created, fill in the policy description carefully, ensuring there are no contradictions. Conflicting rules may confuse the AI and result in inconsistent responses.
17. Editing Existing Policies
To edit a topic or subject, click the pencil icon. To update a description, click into the text box, make your changes, and click Save.
18. Generating Policies and Advanced Settings
If creating a new policy, you may leave the description empty and click Generate Policies to create a baseline set for refinement. In the Advanced tab, you can set response delays and define the minimum confidence score required for automatic messaging.
19. Confidence Scores and Code-Level Instructions
Confidence score details are accessible in the CRM. At the bottom of the Advanced tab, you can add code-level instructions that override policies. Use this feature cautiously, as it has a significant impact on AI behavior.
20. Exploring the Memories Tab
Now, let’s visit the Memories tab. This section displays all AI-saved memories, including memory type, conversation links, descriptions, confidence scores, importance scores, and last update dates.
21. Memories in the CRM
Saved memories are also visible directly within the CRM. The Automation tab here corresponds to the same automation settings found under Guest.
22. Automated Messages Overview
The Automated Messages section—covered in a separate video—allows you to create pre-made messages that are automatically sent to guests based on defined triggers.






















