Axoniq Conference 2025 Day 1 | Query Your Event Store in Seconds: Axoniq Insights Live Demo

Watch Marc Klefter and Stefan unveil Axoniq Insights, a revolutionary analytics platform that lets you query your event store using natural language and SQL—without building projections. See live demonstrations of AI-powered analysis, privacy-preserving queries, and millisecond response times on real event data.

🎯 Featured Speakers:

Marc Klefter - Solutions Architect, Axoniq
Stefan - Software Engineer, Axoniq

🔥 What is Axoniq Insights?

A specialized analytics node for Axon Server that transforms your event store into a queryable data lake:

Ad hoc queries without creating projections
Natural language to SQL powered by AI
Parquet format for optimized column-oriented storage
DuckDB embedded SQL engine
Privacy modes to control data exposure to LLMs
Sub-second query performance on millions of events

💡 Key Features Demonstrated:

Analytics Capabilities:

Query events across multiple Axon contexts
Column-oriented storage for lightning-fast retrieval (11ms queries!)
DuckDB SQL dialect with full query capabilities
Event replication from primary cluster to analytics node
Near real-time data synchronization
Semantic layer for business context

AI-Powered Insights Agent:

Two modes: Privacy (schema only) & Full Access (with data)
Auto-complete with syntax highlighting
Automatic query generation from natural language
Self-debugging and error correction
Multi-step analysis and summarization
Extract JSON/XML data from events
Statistical analysis and reporting

Enterprise Features:

MCP (Model Context Protocol) server for custom agents
Agent-to-agent communication support
Row-based security (coming soon)
Custom LLM provider configuration
GDPR-compliant privacy controls
Integration with PowerBI, Tableau, and BI tools
JDBC/ODBC connectors (planned)

⏱️ Timestamps:

0:00 - Introduction: Event Store Analytics Problem
3:15 - What is Axoniq Insights?
5:40 - Parquet Format & DuckDB Architecture
8:20 - Analytics UI First Look
10:35 - Event Lake Concept & Data Architecture
13:45 - Analytics Node Replication Model
16:30 - Semantic Layer Overview
18:50 - Stefan: Live Demo Begins
21:15 - Auto-complete & Syntax Highlighting
23:40 - Insights Agent Interface (Privacy vs Full Access)
26:10 - Demo: "Which location has most rentals?"
29:35 - Schema Analysis Tool
32:20 - Query Debugging & Self-Correction
35:50 - Demo: "Which days are most popular?"
39:15 - Switching to Full Access Mode
41:30 - Demo: "Analyze bike cancellations by city"
45:25 - Multi-step Analysis & Summarization
49:40 - Customer Journey Analysis
53:15 - Query Performance: 10ms on 100K events
56:30 - Parquet vs Traditional Projections
59:10 - MCP Server for Custom Agents
1:02:45 - Roadmap: BI Tool Connectors (PowerBI, Tableau)
1:05:20 - Security & Privacy Controls
1:08:35 - Ontology & Semantic Layer Deep Dive
1:12:50 - VIP Customer Analysis Example
1:16:30 - Knowledge Graph Potential
1:19:15 - Observability Features Coming
1:21:40 - Integration with Dev Agent
1:23:55 - Early Access & Trial Information
1:25:10 - Q&A: Pricing Model
1:27:30 - Q&A: Payload Types (JSON/XML)
1:29:15 - Q&A: Scalability (Millions to Billions of Events)
1:31:40 - Q&A: Ontology Structure

🎯 Real Demo Use Case: Bike Rental Analysis

Live queries demonstrated:
Which location has the most rentals?
Which days are most popular for bike rentals?
Analyze bike cancellations by city (10% cancellation rate found)
Identify renters with multiple cancellations
Customer journey analysis for potential fraud detection

⚡ Performance Highlights:

8-11 milliseconds for complex aggregations
100,000+ events queried in under 2 seconds
10x-100x faster than traditional projection approaches
Near real-time replication from primary cluster
Column-oriented storage only loads required fields

🔐 Privacy & Security:

Privacy Mode:

✅ LLM sees only schema (no actual data)
✅ GDPR-compliant operations
✅ Safe for sensitive information
✅ Query generation without data exposure

Full Access Mode:

✅ AI sees results for better analysis
✅ Multi-step reasoning
✅ Deeper insights and summaries
✅ Customer journey reconstruction

🏗️ Architecture:

Primary Cluster (Your Apps)

↓ Event Replication

Analytics Node

↓

Parquet Storage (Columnar Format)

↓

DuckDB SQL Engine

↓

Semantic Layer + Ontology

↓

Insights Agent (AI) + MCP Server

↓

Your Queries / Custom Agents / BI Tools

🔗 Resources:

Try Early Access: https://platform.axoniq.io (click "Analyze")
Download Trial: Docker Compose package available
Documentation: https://docs.axoniq.io
Workshop: Hands-on session available at conference