Gtstudio_AiDashboard

AI-powered analytics dashboard for Magento 2. Replaces the stock admin dashboard with KPI cards, charts, and an embedded store assistant — all built from data computed on cron and served from cache.

Preview

AI Studio Ecosystem

Part of the AI Studio suite for Magento 2. See all modules:

Module	Repository	Description
Gtstudio_AiConnector	module-aiconnector	Core AI provider abstraction
Gtstudio_AiAgents	module-ai-agents	Agent & tool orchestration, cron scheduling, execution log
Gtstudio_AiWidgets	module-ai-widgets	Floating admin chat widget + PageBuilder AI generator
Gtstudio_AiDataQuery	module-ai-data-query	Natural-language store analytics (privacy-first)
Gtstudio_AiKnowledgeBase	module-ai-knowledge-base	Document upload & RAG retrieval for agents
Gtstudio_AiDashboard	(this module)	AI-powered KPI dashboard with ML insights

What It Does

• Replaces the default Magento admin dashboard with a richer analytics view
• Metrics are collected on a background cron job and served from a dedicated cache type — no real-time DB load on page load
• KPI cards: today/week/month revenue, orders, new customers, average order value, pending orders
• Charts: 30-day revenue trend (line), orders by status (donut)
• Tables: top 10 products by quantity sold, top 5 customers by LTV, last 10 orders
• Alerts: low-stock product list (threshold configurable)
• AI Insights: one-click AI summary of store performance via the store_assistant agent
• Chat drawer: conversational interface backed by the same agent and the Data Query tool set
• ML features (via php-ml): k-Means customer segmentation (VIP / Active / At-Risk), OLS trend arrows on KPI cards, ±2σ stock depletion anomaly detection

Requirements

• Magento 2.4.4+
• PHP 8.1+
• Gtstudio_AiConnector enabled and configured
• Gtstudio_AiAgents enabled
• Gtstudio_AiDataQuery enabled
• php-ml/php-ml (ML features — pulled automatically by Composer)

Installation

composer require gtstudio/module-ai-dashboard
php bin/magento module:enable Gtstudio_AiDashboard
php bin/magento setup:upgrade

A store_assistant agent record is created automatically via a data patch on setup:upgrade.

Configuration

Stores → Configuration → AI Studio → AI Dashboard

Field	Description
Full Rebuild Schedule	Cron expression for the complete metrics rebuild (default: 5 0 * * *)
Today Refresh Schedule	Cron expression for the incremental today-only refresh (default: 0 * * * *)
Low Stock Threshold	Products with quantity below this value appear in the alerts list (default: 10)
Trend Window (days)	Number of days used for revenue and customer trend charts (default: 30)

How It Works

Data Collection

Heavy queries run in the background on two cron jobs:

Job	Default Schedule	Purpose
BuildDashboardData	5 0 * * *	Full rebuild — all collectors, all time ranges
RefreshTodayData	0 * * * *	Incremental — today's metrics only

Results are written to a dedicated cache type (dashboard_cache_tag) and served as JSON to the frontend.

Cache Keys

Key	TTL	Rebuilt by
dashboard_snapshot_v1	3600 s	BuildDashboardData
dashboard_today_v1	1800 s	RefreshTodayData
dashboard_ai_insights_v1	7200 s	On-demand via Get AI Insights button

To force a rebuild manually:

php bin/magento cache:clean dashboard_cache_tag
php bin/magento cron:run --group=gtstudio_aidashboard

AI Insights & Chat

The Get AI Insights button calls the store_assistant agent with a summary of current metrics and returns a markdown analysis. The chat drawer allows follow-up questions; it has access to all Data Query tools (order_analytics, customer_lifetime_value, product_performance, query_entity).

Extensibility

Adding a custom metric collector

Implement a collector and inject it into DashboardDataService via di.xml:

<type name="Gtstudio\AiDashboard\Model\DashboardDataService">
    <arguments>
        <argument name="collectors" xsi:type="array">
            <item name="my_metric" xsi:type="object">
                Vendor\Module\Model\Collector\MyMetricCollector
            </item>
        </argument>
    </arguments>
</type>

Replacing an ML analyzer

Override via DI preference — for example, to replace the customer segmentation algorithm:

<preference for="Gtstudio\AiDashboard\Model\Analytics\CustomerSegmentationAnalyzer"
            type="Vendor\Module\Model\Analytics\MySegmentationAnalyzer"/>

Swapping the insights agent

Override the block method that resolves the agent code:

<preference for="Gtstudio\AiDashboard\Block\Adminhtml\Dashboard"
            type="Vendor\Module\Block\Adminhtml\Dashboard"/>

Then override getAgentCode() in your subclass to return a different agent code.

ACL Resources

Resource	Controls
Gtstudio_AiDashboard::management	Access to the AI Dashboard page and its AJAX endpoints