Administrator Guide

This guide is intended for Enterprise Administrators who manage Treasure AI Voice deployments across their organization. It covers authentication, device management, user management, and security features.

Before you can access the Admin Dashboard, your organization must be activated. Treasure AI will send you an invitation link — follow the steps in Organization Activation to complete the one-time setup.

The Web Admin Dashboard provides centralized management of your organization's Treasure AI Voice deployment.

• Devices — View and manage all PLAUD devices and their assigned users.
• Recordings — Browse, review, and delete recordings.
• Audit Logs — View a complete audit trail of all activity.
• Members — Manage users — invite, remove, and assign roles.
• AI Lens (optional) — AI-powered analytics and insights.
• Export (optional) — Configure manual and recurring exports of recordings, transcripts, summaries, and device inventory to your Treasure Data CDP.
• Legal — Access Treasure Data's Terms of Service, privacy statement, and other legal documents.

Treasure AI Voice supports two authentication methods:

Direct integrations with SAML 2.0, Microsoft Entra ID (Azure AD), Okta, and other enterprise identity providers are not currently supported. The only built-in social identity provider is Google.

For the time being, organizations that gate Google Workspace sign-in through their corporate IdP can use Sign in with Google to inherit that gate at authentication time. This only affects new sign-ins — it does not terminate active sessions or remove the Treasure AI Voice user record when you deprovision the user upstream. To fully revoke access, an Enterprise Admin must also remove the user from the Members page.

The Devices page (also called Device Fleet) provides a real-time overview of all PLAUD devices deployed across your organization.

At the top of the page, summary cards display:

• Total Devices: Total number of devices registered to your organization.
• Active: Devices that are currently active and operational.
• Locked: Devices that have been remotely locked by an administrator.
• Low Battery: Devices reporting low battery levels.

Devices are grouped by assigned user (name and email). Each device row displays:

If a device is lost or stolen, you can remotely lock it:

1. Navigate to Devices.
2. Locate the device by serial number or assigned user.
3. Click the Lock button in the Actions column.
4. The device status will change to Locked. The action is recorded in the audit log.

The Recordings page displays all transcriptions from organization members.

Use the search bar to find recordings by user email or name. Each recording row displays:

Click a row to open the recording's detail view, where you can view and copy the transcript, summary, action items, and Key Topics (tags). Enterprise Admins can also edit tags directly from this view — see Key Topics (Tags) in the User Guide.

Click any tag chip in the Recordings list to filter it down to recordings carrying that tag. The filter operates across the entire organization's recording set, not just the current page.

The Audit Logs page provides a complete audit trail of all activity on the platform. Logs are retained for 7 years.

Each log entry displays:

Narrow the log using the filters above the table — by date range, event type, and user — so you can answer an audit or compliance question on demand, such as everything a specific user did in a given period, or every occurrence of a single event type across the organization.

The audit log tracks the following events, grouped by category:

The Members page allows you to manage users and team assignments.

Narrow the member list using the filters above the table:

• Days since last login — surfaces members who haven't used the app recently, which is useful for tracking adoption or following up on inactive accounts.
• Status — quickly find users who were invited but haven't yet activated their account.
• Role — Regular User or Enterprise Admin.

1. Click the Invite User button in the top-right corner.
2. Enter the user's email address.
3. Assign a role (Regular User or Enterprise Admin).
4. An invitation email is sent to the user automatically.

AI Lens allows administrators to ask natural language questions about their organization's meeting data, devices, and activity.

Type a question in the text field — for example, "What were the key decisions from last week?" — and click Ask AI. AI Lens analyzes your organization's recordings and returns an executive summary with insights.

The analysis scope is displayed below the search field (e.g., "Analyzed: 75 recordings, 11 users, 15 devices, 500 events (last 30 days)").

Click Configure in the top-right corner to adjust analysis settings.

The Export page lets Enterprise Administrators send recordings, transcript segments, summaries, and device inventory to a Treasure Data CDP database. The data can then be joined with other customer data for analytics, segmentation, and activation.

Treasure AI Voice supports both manual and recurring exports. Exports run in the background, so the console returns control after creating an export run. Use the History tab to monitor the run and review its result.

Before configuring an export, make sure you have:

• An Enterprise Administrator account in Treasure AI Voice.
• CDP export enabled for your organization.
• A destination database in the Japan (AP01) region of Treasure Data CDP.
• A dedicated Treasure Data CDP API key with Import Only or General Access permission on the destination database.

CDP export is disabled by default. To enable it:

1. Sign in to the AI Voice Console as an Enterprise Administrator.
2. Open the Export page.
3. Select Treasure Data CDP.
4. Click Enable export.

If the option is unavailable, contact your Treasure AI representative.

Save the destination and credential before running a manual or recurring export:

1. Open the Settings tab on the Export page.
2. Enter the destination Database.
3. Enter a dedicated Write-only API Key.
4. Optionally expand Advanced and change the Recordings table and Transcript segments table names.
5. Click Save settings.

The database and table settings are restored when you return to the Export page. A saved API key is shown only as a masked value. The full key is not returned to the console, included in an export run payload, or written to the audit log.

To replace the credential, click Replace and save a new key. To remove it, click Remove. Removing the saved key disables an enabled recurring schedule.

The destination database name must be 3–64 characters, begin with a lowercase letter, use only lowercase letters, digits, or underscores, and must not be export. Table names must be 3–128 characters and use only lowercase letters, digits, or underscores.

1. Confirm that the destination database and saved API key are shown on the Settings tab.
2. Optionally expand Advanced and set a From and To date range.
3. Optionally select Re-export previously exported recordings.
4. Click Run export now.
5. Open the History tab to monitor the run.

The date range filters recordings by their recorded date. The To date includes the entire day. The range applies to recordings, transcript segments, and summaries. It does not filter devices by recording date.

A manual export does not need to remain open in the browser. The export continues in the background after the run starts.

Recurring exports use the saved destination, table settings, and write-only API key. Save and verify those settings before enabling a schedule.

1. Open the Schedule tab.
2. Turn on Scheduled export.
3. Select an Interval:
   • Every hour
   • Every 6 hours
   • Every 12 hours
   • Every 24 hours
   • Weekly
4. Select the behavior for the first scheduled run.
5. Save the schedule.

The shortest available interval is one hour. AI Voice CDP export uses the listed intervals; it does not provide a custom cron expression. Saving an enabled schedule starts an export immediately. Subsequent exports run at the selected interval from that initial run. For example, if an hourly schedule is saved at 10:23, the initial export starts immediately, followed by scheduled exports at 11:23, 12:23, and so on.

A manual or scheduled export cannot start while another export is running for the same organization. A recurring schedule is configured at the organization level and uses the saved credential without requiring an administrator to be signed in when the run starts.

Select one of the following options when enabling a schedule:

• Only recordings from now on: Starts the export boundary when the schedule is enabled. Historical recordings are not backfilled automatically, but records updated after that point are included.
• Send everything not yet exported first: Starts with the organization's existing exportable data and then continues with incremental exports. A large recording history may require multiple runs.

The History tab lists recent manual and scheduled export runs. Each row shows:

• Start time
• Trigger type
• Status
• Accepted recording rows
• Accepted transcript segment rows
• Rejected rows
• Duration
• Run ID

Select a row to view the destination database, accepted and rejected counts by table, failure information, and recordings rejected because their event was too large.

Invalid credentials and insufficient Treasure Data CDP database permissions can appear as a failed run because the export runs asynchronously. Grant the key's user Import Only or General Access permission on the destination database, then run the export again.

A recording whose serialized event exceeds the 1 MB per-event limit is listed in the run details. The recording row is rejected, but its transcript segments and summary may still be exported. The recording is retried by later runs for a limited number of attempts. The console does not show the serialized event size before export. If the limit is exceeded, review the rejected recording in the run details.

Each exported event includes the following fields:

• unique_id: A stable identifier for the source record.
• updated_at: The source record's last-updated timestamp, in Unix epoch seconds.
• time: The timestamp used for Treasure Data time partitioning.

time and updated_at have different purposes. Use time for partition filtering and updated_at to determine which copy of a record is current.

Scheduled and undated exports use an incremental checkpoint. Recordings and devices updated at or after that checkpoint are included in the next run. A date range narrows recording candidates by recording date while preserving the incremental check.

Treasure Data CDP appends incoming events instead of replacing existing rows. Repeated exports can therefore contain multiple rows with the same unique_id. Deduplicate downstream by unique_id, keeping the row with the greatest updated_at.

Every export run is recorded in Audit Logs under the TD CDP Export event type. Run status and row counts are available in History and the run details, not in Audit Logs. Changes to the destination settings and saved credential are also recorded. The API key itself is never written to the audit log.

• The minimum recurring-export interval is one hour.
• Only one export can run at a time for an organization.
• A single recording event cannot exceed 1 MB.
• A batch cannot exceed 4 MB.
• A run can finish with Partial when it reaches the worker time limit or when Treasure Data rejects rows.
• If a run fails, review its failure information in History and verify the destination database, API key, and database permission.
• If a run reports Nothing to send, verify the date range and whether previously exported recordings should be included.

For the exported table schemas, data relationships, and querying examples, see Treasure Data CDP Integration.

The Legal page provides direct links to Treasure Data's legal documents that govern the use of Treasure AI Voice:

• Terms of Service
• Voice Supplemental Terms
• AI Terms
• AI Acceptable Use Policy
• Privacy Statement

Each link opens the current version of the document on the Treasure Data website.