bchic Analytics MCP Server

With our MCP server, you ask: “Show me top performers by channel and region.” Seconds later, you have the analysis, including visualization and recommendations for action.

It's not hype. That is infrastructure.

‍

How it works

We follow the Authenticated Remote MCP spec. The server is centrally hosted and managed.

Tools available:

• query_metrics — Query any dimensions and metrics from your analytics data
• detect_anomalies — Automatic detection of conversion vulnerabilities and performance issues
• analyze_intent — User intent analysis based on journey data
• generate_dashboard — Create dashboards from natural language or screenshots (Q2 2026)

More functionality is constantly being added.

‍

The technical difference

Other analytics platforms are experimenting with AI features on cookie-based systems. The problem: The data architecture is fundamentally wrong for AI.

bchic Analytics is built from the ground up to be cookie-free:

• No personal data — MCP accesses aggregated analytics data, not user profiles
• No consent complexity — No PII & cookies = no banners = no consent requirement
• GDPR-compliant by design — The architecture makes data breaches technically impossible

That means: Full AI power without data protection overhead. That is the difference between “we also have AI” and “our AI is built differently because our data is different.”

‍

From hours to seconds

In our benchmarks, we reduce analyses that take several hours on other platforms to seconds.

Example: Multi-source analysis

Traditional workflow:

1. Exporting GA4 metrics
2. Drag CRM data
3. Merge into Excel
4. Build pivot tables
5. Create charts
6. Present in PowerPoint

Time: 2-4 hours

With MCP:

“Show me the most profitable customer channels by product category and region”


The AI:

• Detects intent (profitability analysis by acquisition channel)
• Identifies required data sources
• Pulls data from analytics infrastructure
• Performs analysis
• Create visualization
• Provides recommendations for action

Time: seconds.

It's not demo magic. That is production.

‍

Dashboard generation from screenshots

Q2 2026 Feature: The AI can create dashboards in bchic Analytics — from natural language descriptions or screenshots of GA4, Power BI, Tableau.

Do you have a perfectly configured GA4 dashboard? Upload screenshot, say “Build this in bchic Analytics.” It's done.

This works because MCP not only moves data, but also understands semantics. AI knows your metrics, dimensions, business logic. It doesn't copy — it adapts intelligently.

‍

Early Access

Q1 2026 gives the first users access. This includes:

• Natural language queries across the entire analytics infrastructure
• intent-based vulnerability analysis
• Automated Anomaly Detection
• Report generation

Q2 2026 full rollout with dashboard generation of text and screenshots.

‍

The technical architecture

MCP defines three roles:

• Host — The app that the user sees (Claude, Cursor, etc.)
• Client — Manages connections and orchestrates request/response
• servers — Where your data lives (bchic Analytics)

The request/response lifecycle:

1. Tool Discovery — Client asks server for available capabilities
2. Intent Recognition — AI analyses queries and determines required tools
3. Execution — Tools are called on server (s) (possible in parallel)
4. Contextualization — AI processes raw data into insights
5. Response — User receives comprehensible answer with context

Model-agnostic by design. You can switch between Claude, GPT-4, Gemini without rebuilding integrations. Zero vendor lock-in.

‍

Why cookie-free is the game changer

Most analytics platforms:

• Cookie-based tracking
• personal data
• Consent banner required
• 50% + data loss due to rejections
• GDPR compliance overhead

bchic Analytics:

• Cookie-free server-side tracking
• No personal data
• No banners
• 100% data depth
• GDPR-compliant by architecture

This is not a trade-off between AI power and data protection. That is both.

‍

roadmap

Q1 2026 (Early Access)

• Natural Language Queries
• Anomaly Detection
• Intent Analysis
• Automated Reports

Q2 2026 (full rollout)

• Natural language dashboard generation
• Dashboard generation from screenshots
• Multi-source analytics (bchic + external data sources)
• Real-Time Collaboration Features

Q3+ 2026

• Predictive analytics
• Custom tool creation (build your own MCP tools for specific workflows)
• Advanced Intent Modeling

‍

Props to Anthropic and Hetzner for the infrastructure.