Insights

Insights is where AI4Love's intelligence becomes visible. Seven specialized AI agents analyze your supporter data nightly, and the Insights dashboard surfaces their findings — patterns, at-risk signals, conversion opportunities, and campaign intelligence — so your team knows where to focus.

What Insights Surfaces

Each insight is generated through a two-step process:

• Deterministic Pattern Detection - Rules-based analysis identifies eligible supporters. The AI does not select who to analyze — patterns and thresholds do.
• Narrative Generation - Claude AI receives the detected pattern and supporter context, then generates a human-readable insight with a recommended action.

Every insight includes the pattern that triggered it and the data behind it, so your team always knows why a supporter was flagged and what action is recommended.

The Seven Agents

• At-Risk Detection - Identifies supporters showing signs of declining engagement — fewer donations, missed events, reduced volunteer hours.
• Conversion Opportunities - Finds supporters ready for deeper engagement — volunteers who might become donors, one-time donors who could become recurring.
• Recognition Triggers - Detects milestones worth celebrating — giving anniversaries, volunteer hour thresholds, consistent attendance streaks.
• Campaign Intelligence - Surfaces individual-level campaign pattern signals to help personalize outreach timing and messaging.
• Relationship Deepening - Identifies opportunities to strengthen existing connections through shared interests, event invitations, or personal touchpoints.
• Cross-Agent Intelligence - Synthesizes patterns across all other agents to surface meta-insights and cross-cutting themes.
• KindMind Enrichment - Draws on your organization's uploaded knowledge base to enrich agent insights with organizational context.

Intelligence Layer, Not a Campaign Tool

AI4Love surfaces recommendations. Humans decide what to do with them.

• No Automated Outreach - AI4Love never sends emails, makes calls, or contacts supporters directly
• Staff Review Required - Every insight is reviewed by your team before any action is taken
• Recommendations, Not Automation - The dashboard provides suggested actions and draft messages — your team decides what ships
• Transparent Reasoning - Every insight shows the triggering pattern and supporting data

Knowledge Base Integration

Make your insights smarter with your organization's own context:

• Document Upload - Upload PDFs, Word documents, and text files to build your knowledge base
• Semantic Search - Vector-powered search finds relevant context from your documents automatically
• Context Enrichment - The KindMind agent uses your knowledge base to ground insights in your organization's language and priorities
• Coverage Tracking - Monitor what your knowledge base covers and where gaps exist

How It All Connects

Agents run nightly, orchestrated by Make.com:

• New activity flows into the Participation table and is automatically queued for agent processing
• Each agent analyzes queued records and writes insights back to Airtable
• Insights include a headline, pattern type, confidence level, associated supporter, and recommended action
• The Insights dashboard surfaces these across filterable views so your team sees what matters most