Data Management
The Data Management page in Blockli Assistant (found under Settings → Data Management) shows you the list of scheduled data sync actions that keep your AI agent informed about your WordPress site. These recurring background jobs collect structured site data and securely transmit it to the Blockli middleware, where the AI agent can use it for reasoning — without ever directly accessing your WordPress database.
Blockli AI agents operate in a sandboxed environment completely separate from your WordPress installation. This means the AI never reads from or writes to your database directly. Instead, a curated summary of your site's activity and structure is synced on a daily schedule so the agent has the context it needs to provide accurate, relevant responses and insights.
This sandbox architecture provides two key benefits:
• Security — Your WordPress database is never exposed to the AI processing layer.
• Data Protection — Only precomputed analysis data is used for agent reasoning, reducing exposure of raw data.
Blockli syncs the following data types, depending on which plugins are active on your site:
Activity counts, group memberships, engagement metrics, and aggregated community analytics such as average activity per user, activity type breakdowns, and top contributor summaries.
Device platform distribution, app version usage, registration trends, and active device counts. This helps the AI understand how your community accesses the platform.
Post metadata including titles, types, statuses, publish dates, comment counts, and taxonomy assignments. Post body content is not synced — only structural metadata.
Group names, member counts, activity levels, engagement scores, and growth metrics. This gives the AI context about community group health and participation patterns.
Forum structure, topic counts, reply activity, engagement rates, and author activity summaries. Forum post content is not synced — only metadata and analytics.
Order summaries, revenue analytics, product performance, category breakdowns, payment method distribution, and customer purchase patterns. Customer email addresses are excluded by default.
Blockli Assistant implements multiple layers of privacy and security for synced data:
All data sync requests are transmitted over HTTPS/TLS to the Blockli middleware endpoint, ensuring data is encrypted during transmission and cannot be intercepted.
Blockli includes an encryption layer using RSA and AES-256-CBC. A random secret key is generated for each encryption operation, encrypted with the Blockli public key using RSA-OAEP, and used to encrypt the payload with AES-256-CBC. Data stored on the middleware side is protected with encryption at rest.
The sync process is designed to minimize PII exposure. Customer email addresses are excluded by default (controlled via a WordPress filter). Post content and message bodies are never synced — only metadata and aggregated metrics. The AI agent uses precomputed analysis data for its reasoning, not raw personal records.
The AI agent runs in a completely isolated environment with no direct access to your WordPress database, files, or server. It can only reference the structured, pre-synced data summaries — it cannot query, modify, or extract data from your site on its own.
The Data Management page displays a table of all scheduled sync actions with the following information:
• Sent (green) — The sync completed successfully.
• Pending (yellow) — The sync is scheduled and waiting to run.
• Failed (red) — The sync encountered an error.
• Event — The name of the data sync action (e.g., "User Insights Data").
• Status — Current state of the action.
• Recurrence — How often the action runs (default: every 1 day).
• Scheduled date — When the action is next scheduled to execute.
• Action — Run or cancel the action manually.
You can also use the top metrics bar to quickly see how many actions are in each state (All, Sent, Failed, Pending), filter actions by status, search by event name, and use bulk actions to delete selected items.
All data sync actions run automatically once every 24 hours using the Action Scheduler library (with WP-Cron as a fallback). Data is processed in batches to avoid overloading your server — for example, user data is processed in batches of 100, and WooCommerce customer data in batches of 50.
If your site is disconnected (no tenant ID configured), all scheduled sync actions are automatically cleared to prevent unnecessary processing.