How can we help? 👋

Setting Up a Maintenance AI Agent

A detailed look at how to set up your maintenance AI agent.

A Maintenance AI Agent can answer maintenance-related calls, help diagnose common issues, collect important details from residents, and submit maintenance requests automatically. This allows your team to capture requests efficiently while providing residents with immediate assistance.


Before You Begin

To create and use an AI agent, you must first have a Twilio phone line connected.

If you plan to use text messaging with the AI agent, your Twilio number must also be registered for A2P 10DLC.

Important:

  • A2P 10DLC registration is only required for texting.
  • If your AI agent will only handle calls, registration is not required.

Creating a New AI Agent

To create a new AI agent:

  1. Navigate to AI Agents.
  1. Click New Agent at the bottom of the page.
  1. Configure the agent settings described below.

Agent Profile

The Agent Profile identifies your AI agent in the system. Many of these fields are internal but help your team manage multiple agents.

Notion image

Fields include:

Setting
Description
Avatar
Upload an image to easily identify the agent in the agent list.
Agent Name
The name the AI agent will use when answering calls.
Title
Internal title describing the agent’s purpose.
Description
Internal description to help your team understand what the agent handles.

Outward-Facing Voice Settings

These settings control how the AI agent sounds during calls.

Notion image
Setting
Description
Call Voice
Select the voice the AI agent will use. Use the speaker icon to preview available voices.
Voice Speed
Adjust how fast the agent speaks. This may need to be fine-tuned after listening to real calls.
Call Responsiveness
Controls how quickly the agent responds after a caller speaks. Options include Patient, Balanced, and Responsive.
Call Tone and Attitude
Defines how the agent communicates with callers. Options include Professional, Casual, Robotic, and Empathetic.

Tip: The Responsive option replies quickly but may occasionally interrupt callers.


Email Notifications

You can configure email alerts for certain call situations.

Notion image
Notification
Description
Off-Topic Email Notification
Sends an email if a caller asks something unrelated to the agent’s role (for example, a leasing question on a maintenance line).
Skip Agent Notification
Sends an email if a caller bypasses the AI agent.
Emergency Call Notifications
Sends an email if the caller says their issue is an emergency.

Notifications can be sent to:

  • Any user in the system
  • Any external email address

Agent Limits

Agent limits allow you to control how many calls or minutes an AI agent can handle each month.

Notion image
Limit Type
Description
Call Count
Maximum number of calls the AI agent can handle.
Call Minutes
Maximum total call time the agent can handle.

Important details:

  • Limits reset monthly.
  • If both limits are enabled, the first limit reached will stop the agent from answering additional calls.
  • If the limit is reached during an active call, the agent will finish the call but will not accept new calls afterward.

When limits are reached, calls will route to the Failover Number configured later.


Agent Metrics

The Agent Metrics section displays performance data for the agent.

Notion image

Metrics include:

  • Number of calls handled
  • Total call duration
  • Number of Cards created

These metrics are shown for the rolling last 30 days.


Agent Role: Maintenance Operations

When setting up a maintenance AI agent, select the Maintenance Operations role.

Notion image

This role includes a built-in prompt that instructs the AI agent to:

  • Diagnose common maintenance issues
  • Ask step-by-step troubleshooting questions
  • Submit maintenance requests (when enabled through capabilities)
  • Transfer calls to management if transfer capabilities are configured
  • Log non-maintenance issues for staff follow-up

The maintenance agent communicates conversationally by:

  • Asking one question at a time
  • Confirming details with the caller
  • Maintaining an empathetic and helpful tone

The role also includes preloaded troubleshooting guidance for common maintenance issues.


Optional Overrides

You can customize how the AI agent introduces itself or behaves.

Notion image

Override Agent Identity

This allows you to modify the AI agent’s role description or responsibilities.

Use this option if you want the agent to represent your organization differently.

Override Agent Greeting

A default greeting is included with the maintenance role, but you can replace it with a custom message to match your company’s branding or tone.

Example:

“Thank you for calling maintenance at Green Valley Apartments. I can help collect details about your request and assist with troubleshooting.”


Data Collection Settings

The maintenance AI agent automatically collects important details from the caller.

Default collected information includes:

  • Contact name
  • Phone number
  • Unit address
  • Incident location
  • Preferred service time
  • Issue type
  • Issue details
  • Issue start date
  • Whether the issue occurred before
  • Troubleshooting steps taken
  • Potential damage risk
  • Emergency status
  • Request priority
  • Permission to enter
  • Pet information
  • Property access instructions

Custom Data Collection

You can enable Custom Data Collection to control which fields the AI agent collects.

Notion image

Options include:

  • Turning default fields on or off
  • Selecting the field to map to
  • Adding custom data fields

Adding Custom Fields

When creating a custom field, you must provide:

Field
Description
Field Label
The name of the field
Question
The question the AI agent asks the caller
Description
Information explaining the field’s purpose

If the board where cards are created already has a field with the same name, the collected data will populate that field.

If the field does not exist, the system will automatically create a new field on the card.


Knowledge

Notion image

The Knowledge section allows you to provide additional information the AI agent can reference during calls.

Relevant Locations

You can optionally add information about specific properties or units.

Only locations marked as published for rent in listing settings will appear as options here.

This allows the AI agent to reference property-specific information when assisting callers.

  • Allows you the option to create multiple agents to service specific locations or properties (this is a rare instance, best practice is to have one agent handle all of your properties)

Agent Knowledge

This information should help the AI agent answer questions about your organization.

Examples include:

  • Maintenance office hours
  • Emergency maintenance policies
  • Expected repair timelines
  • Maintenance request procedures

Important:

The knowledge section is not used to instruct the AI agent how to perform actions. It is only used to provide information the agent can share with callers.


Triggers

Triggers determine how the AI agent answers calls.

Notion image

For maintenance agents, the trigger is typically:

Inbound Call

To configure the trigger:

  1. Select the Twilio phone line
  1. Assign it to the AI Agent

Failover Number

The Failover Number acts as a backup if the AI agent cannot answer a call.

This may occur if:

  • Agent limits are reached
  • The agent is unavailable
  • The system cannot process the call

When this happens, the call will automatically route to the failover phone number you configure.

Other Triggers:

  • Single Number: Will bypass your AI Agent and go to an external number.
  • Multiple Numbers: Will bypass your AI Agent and ring to multiple external numbers.
  • Ring Schedule: Will allow a schedule to be set, for example you can have your AI Leasing Agent answer calls outside of business hours and a live agent answer during business hours.

Capabilities

Capabilities define what actions the AI agent can take during or after a call. While the maintenance agent will always answer and guide callers through troubleshooting and information collection, capabilities determine how that information is used and what the agent can do next.

Notion image

You can enable one or multiple capabilities depending on your maintenance workflow.

Examples include:

  • Creating or updating cards
  • Submitting maintenance requests
  • Transferring calls to staff
  • Logging requests for follow-up

These capabilities define the actions the AI agent takes after collecting information from the caller.

Create or Update Card

This capability allows the AI agent to create a new card or update an existing card based on information collected during the call.

When configuring this capability:

  1. Select the board where cards should be created.
  1. Choose the stage where new cards will appear.

For maintenance workflows, this is commonly configured on a Work Orders board, but any board can be selected.

The system will match cards using the contact name and phone number. If a matching card already exists based on the board’s deduplication rules, the card will be updated instead of creating a new one.

Example Use Case

A resident calls to report a leaking sink.

The AI agent collects the maintenance details and automatically creates a work order card in the appropriate stage.


Transfer Call

The Transfer Call capability allows the AI agent to transfer callers to a team member or phone number when certain conditions are met.

Each transfer destination requires its own transfer rule.

For example, if you want the agent to transfer calls to multiple people, you will need to create separate transfer rules for each destination.

Example setup:

Destination
Phone Number
Maintenance Supervisor
(555) 555-1111
Property Manager
(555) 555-2222
Emergency Maintenance Line
(555) 555-3333

Within each transfer rule, you can configure:

  • Conditions that determine when the transfer should occur
  • The message the AI agent will say before transferring the call

Example transfer message:

“This issue may require immediate assistance. Let me transfer you to a member of our maintenance team.”

Card Lookup

The Card Lookup capability allows the AI agent to retrieve and share information from an existing card.

This is commonly used for checking the status of maintenance requests.

To configure this capability:

  1. Enable Card Lookup
  1. Select the Work Orders board

The system includes a preconfigured conversation model and standard fields designed for maintenance workflows.

If needed, you can customize the configuration by selecting specific fields you want the AI agent to read back to callers.

Example information the agent could provide:

  • Work order status
  • Scheduled service time
  • Assigned technician
  • Request details

Example Use Case

A resident calls to check the status of a previously submitted maintenance request.

The AI agent looks up the work order and provides the current status.


End Call

The End Call capability allows the AI agent to end a call when specific conditions occur.

This capability is not commonly used, but it can be helpful in situations where the AI agent and caller become stuck in a conversational loop.

For example:

  • The caller attempts to end the conversation.
  • The AI agent says goodbye.
  • The caller continues asking additional questions.
  • The conversation repeats indefinitely.

In these cases, the End Call capability ensures the call is properly closed.

This feature is typically only enabled when troubleshooting looping conversations.


Finalizing Your Agent Setup

Once all settings are configured:

  1. Review your agent profile
  1. Confirm voice and behavior settings
  1. Verify data collection and capabilities
  1. Click Save

Your maintenance AI agent will then be ready to begin answering calls.



Next Step:

Once your agent profile, triggers, and capabilities are configured, save and activate the agent to begin handling maintenance calls.


Example Maintenance AI Agent Flow

When a resident calls to report a maintenance issue, the AI agent answers the call and begins guiding the resident through the reporting process.

During the conversation, the agent will:

  • Collect the resident’s name
  • Confirm the property or unit address
  • Ask questions to intake details about the issue
  • Provide basic troubleshooting guidance when applicable

Once the necessary information has been collected, the AI agent will automatically create a new card on your Work Orders board in the Requested stage.

Notion image

If desired, you can create a dedicated stage for these requests, such as AI Created Tickets, to make them easier to identify and review.

If you are using Rentvine, you can also create this stage within your PMS so it will sync directly to Aptly.

After the request is submitted, the AI agent can send the resident a confirmation text message (requires a Twilio line registered for A2P 10DLC). The message will include the work order number and confirmation that the request has been received and is being processed.

All related activity—including created cards, call transcripts, and text messages—can be reviewed in the Agent Activity Log under Activities. This allows your team to quickly review the conversation and follow up on the request if needed.

Notion image

Related Articles

 
Did this answer your question?
😞
😐
🤩