GitHub - jonasarmbrust/AIStrategyHub: Self-evolving AI maturity platform — an autonomous Evolution Agent researches new standards, proposes checkpoints, and grows your framework from 101 to 190+. Powered by Gemini AI, 6 global standards, and agentic AI.

Is your organization truly ready for AI? Find out — for free.
An open-source platform that helps you assess, plan, and track your AI strategy
using best practices from 6 global standards — powered by Google Gemini AI.
And it evolves itself.

🚀 Try the Live Demo →
_{Hosted on Google Cloud Run — protected instance. Request access or contact me for a demo key.}

💡 What Is This?

In one sentence: AI Strategy Hub is a free, self-evolving tool that tells you how ready your organization is for AI — and autonomously keeps your assessment framework current.

The Problem

Every company knows they should be using AI. But most don't know where to start, what they're missing, or how mature their current approach actually is. Meanwhile, dozens of excellent guides exist — from the EU AI Act to Google's AI Adoption Framework — but they each have different formats, focus areas, and languages. Reading them all would take weeks.

And hiring a consultancy? That's costly, slow, and gives you a one-time snapshot that's outdated within months — especially in a field where new regulations, tools, and best practices emerge almost weekly.

Traditional maturity models have an even deeper flaw: they're frozen the moment they're published. The AI landscape evolves weekly; your assessment framework should too.

The Solution

AI Strategy Hub merges 6 internationally recognized standards (NIST, EU AI Act, Google, Microsoft, OWASP, UNESCO) into a living meta-model across 7 key areas — and then evolves it autonomously. You can:

🧬 Let the AI evolve your framework — The Evolution Agent autonomously researches new standards, generates checkpoint proposals, detects redundancies, and integrates approved changes on a configurable schedule. The framework has already grown from 101 → 190+ checkpoints this way.
✅ Self-assess — Walk through an interactive checklist and get your maturity score instantly
📄 Upload existing strategy documents — The AI reads them and evaluates your readiness automatically
📋 Follow the AI Playbook — A phased implementation roadmap (Foundation → Transformation) with effort estimates and priority badges
🧠 Ask the AI Advisor — "What should we focus on first?" — and get answers based on your specific gaps
📈 Track progress over time — Every assessment becomes a snapshot, so you can see how you improve
🔬 Stay current — automatically — The Evolution Agent scans the web for new frameworks and regulations, so you never fall behind

Think of it as a fitness tracker for your AI strategy — instead of steps and heart rate, it measures governance, data readiness, talent, and ethics. And unlike a static report, it gets smarter over time.

Tip

The AI world moves fast. Your assessment tool should too. Traditional maturity models are frozen the moment they're published. AI Strategy Hub's Evolution Agent autonomously discovers, evaluates, and proposes new checkpoints — keeping your framework as current as the field itself.

Who Is This For?

Role	What you get
CTO / Head of AI	A clear picture of organizational AI readiness with concrete action items
IT Project Manager	A structured framework to plan and prioritize AI initiatives
Compliance / Legal	EU AI Act gap analysis with article-level mapping and risk scores
Consultant / Advisor	A reusable assessment tool for client engagements — free and white-label-ready
Student / Researcher	A hands-on way to learn about AI governance frameworks

🧬 The Evolution Strategy — Agentic Framework Evolution

Evolution Dashboard — Monitor, approve, and manage how your AI maturity model grows and refines itself over time

This is what makes AI Strategy Hub fundamentally different from every other maturity model.

Traditional assessment frameworks are static documents — published once, then frozen. AI Strategy Hub's Evolution Engine is an autonomous agent that continuously improves the underlying framework:

How It Works

┌─────────────────────────────────────────────────────────────────────┐
│                    🧬 EVOLUTION CYCLE                               │
│                                                                     │
│   ┌──────────┐    ┌──────────┐    ┌──────────┐    ┌──────────┐     │
│   │ Research  │───▶│ Evaluate │───▶│ Propose  │───▶│ Dedupe   │     │
│   │  Scan    │    │ (Gemini) │    │Checkpoints│   │(Embeddings)│    │
│   └──────────┘    └──────────┘    └──────────┘    └──────────┘     │
│        │                                               │            │
│        │              ┌──────────┐                     │            │
│        │              │  Human   │◀────────────────────┘            │
│        │              │  Review  │                                  │
│        │              └────┬─────┘                                  │
│        │                   │ Approve / Reject                       │
│        │              ┌────▼─────┐                                  │
│        │              │Integrate │                                  │
│        │              │  & Lock  │                                  │
│        │              └────┬─────┘                                  │
│        │                   │                                        │
│        ▼                   ▼                                        │
│   ┌────────────────────────────────────┐                            │
│   │ 📊 Living Meta-Model (190+ CPs)   │                            │
│   │    Snapshots · Rollback · Growth   │                            │
│   └────────────────────────────────────┘                            │
└─────────────────────────────────────────────────────────────────────┘

🔬 Automated Research Scanning — The Evolution Agent triggers Tavily-powered web research across all 7 dimensions, discovering new AI governance frameworks, regulations, and academic papers.
🧠 AI-Powered Evaluation — Each research result is evaluated by Gemini against all existing checkpoints using multi-criteria assessment (quality, impact, novelty).
📝 Checkpoint Proposal Generation — Gemini generates concrete new checkpoint proposals with dimension assignment, maturity level, source attribution, and quality scores.
🔍 Embedding-Based Deduplication — A Redundancy Scanner uses Gemini embeddings + cosine similarity to detect semantically duplicate checkpoints and flags them for merging.
👤 Human-in-the-Loop Review — All proposals are queued in the Evolution Dashboard for human approval or rejection — full transparency, full control.
🔒 Atomic Integration — Approved checkpoints are merged into the living meta-model with CrossProcessFileLock for safe concurrent writes.
⏰ Scheduled Autonomy — Configurable cron-like scheduling (weekly by default) — the framework evolves on autopilot while you review.

Evolution Dashboard Features

Feature	Description
Live Status Bar	Real-time evolution status with countdown to next scheduled run
KPI Cards	Total runs, checkpoints integrated, redundancies resolved, avg quality score
Evolution Timeline	Chronological log of all evolution runs with expandable details
Framework Growth Chart	Stacked visualization of checkpoint count growth per dimension
Pending Proposals Queue	AI-generated proposals with quality/impact/novelty scores — approve or reject
Redundancy Scanner	Side-by-side comparison of semantically similar checkpoints with similarity %
Snapshot & Rollback	Framework versioning with one-click rollback to any previous state
Configuration Panel	Schedule frequency, auto-approve toggle, quality thresholds

Important

From 101 to 190+ checkpoints — autonomously. The Evolution Engine has already nearly doubled the framework's coverage by discovering and integrating checkpoints from sources like arXiv papers, MIT Sloan Management Review, appliedAI, and Google's Data & AI Strategy Assessment.

✨ Key Features

🧬 Evolution Agent `NEW`

Autonomous framework evolution — researches new standards, generates checkpoint proposals, detects redundancies, integrates approved changes on a configurable schedule. The framework evolves itself.

📋 AI Strategy Playbook `NEW`

Phased implementation roadmap (Foundation → Experimentation → Operationalization → Scaling → Transformation) with effort estimates, priority badges, and per-dimension progress tracking.

🔬 AI Deep Dives `NEW`

Per-checkpoint detailed analysis powered by Gemini. In-depth implementation guidance and best practices — connected from EU AI Act gaps ("Fix via AI Deep Dive").

🧠 AI Strategy Advisor

Interactive chatbot powered by Gemini 3.1 Pro — knows your assessment scores, gaps, and research sources. Provides context-aware strategic advice.

📊 Maturity Assessment

Interactive checklist across 7 weighted dimensions with automated scoring, radar chart visualization, and level classification (1–5).

🔍 Document Analyzer (RAG)

Drag & drop your AI strategy docs (PDF, DOCX, TXT). The RAG pipeline uses embeddings + Gemini to evaluate all checkpoints with confidence scoring.

🔗 Evidence Chain

Full traceability — click any checkpoint to see the AI's reasoning, evidence text, confidence %, and the original source chunks.

⚖️ EU AI Act Compliance

Maps gaps directly to EU AI Act requirements with compliance readiness score, regulatory exposure level, and article-level fine amounts.

🎮 Gap Simulator

"What-If Analysis" — toggle checkpoints and see real-time impact on your maturity score. Discover highest-ROI actions.

🔬 Research Agent

Automated web research via Tavily API. Discovers new frameworks & regulations, evaluates relevance with Gemini. Feeds directly into the Evolution Engine.

🏗️ Framework Builder

Extract novel checkpoints from research documents and integrate them into the living meta-model. The framework evolves.

🕐 Assessment Timeline

Track your AI maturity over time. Every assessment is saved as a snapshot — the dashboard shows score progression with deltas and visual comparison.

🗺️ Strategic Roadmap

AI-generated prioritized action plan with effort estimates, quick wins, and milestone recommendations.

🌍 Fully Bilingual (EN/DE)

Every UI element, report, and AI response is available in English and German — toggle with one click.

📸 Screenshots

Dashboard — Your AI maturity at a glance with radar chart and dimension scores

Evolution Dashboard	AI Strategy Playbook

Autonomous framework evolution — monitor runs, approve proposals, track growth	Phased implementation guide from Foundation to Transformation

Document Analyzer (RAG)	AI Strategy Advisor

Upload strategy docs — AI evaluates every checkpoint with strengths & gaps	Context-aware AI consultant powered by Gemini

EU AI Act Compliance	Gap Simulator

Regulatory gap analysis with article-level compliance mapping & fines	What-If analysis — toggle checkpoints and see real-time score impact

Dependency Map	Framework Explorer

Interactive force-directed graph of checkpoint relationships	Browse all checkpoints with source traceability and AI Deep Dives

🏗️ Architecture

graph TB
    subgraph Frontend["🖥️ Frontend — Vite + Vanilla JS"]
        FE_PAGES["16 SPA Pages<br/>Dashboard · Assessment · Analyzer<br/>Advisor · Simulator · Roadmap<br/>Evolution · Playbook · Deep Dive · ..."]
        FE_I18N["i18n (EN/DE)"]
        FE_SANITIZE["DOMPurify XSS Protection"]
    end

    subgraph Backend["⚙️ Backend — FastAPI + Python"]
        ROUTES["12 API Routers<br/>50+ Endpoints"]
        AUTH["Auth Middleware<br/>API Key + Rate Limiting"]
        
        subgraph AI["🧠 AI Layer"]
            GEMINI_PRO["Gemini 3.1 Pro<br/>Advisor · Summaries · Evolution"]
            GEMINI_FLASH["Gemini 2.5 Flash<br/>Batch Evaluation"]
            GEMINI_EMBED["Gemini Embeddings<br/>Document Vectors · Redundancy Detection"]
        end

        subgraph Evolution["🧬 Evolution Engine"]
            EVO_AGENT["Evolution Agent<br/>Orchestrator"]
            EVO_EXTRACT["Checkpoint Extractor<br/>Gemini-powered"]
            EVO_REDUNDANCY["Redundancy Detector<br/>Embedding Similarity"]
            EVO_SCANNER["Research Scanner<br/>Tavily Deep Scan"]
            EVO_SNAPSHOT["Snapshot Manager<br/>Versioning & Rollback"]
            EVO_SCHEDULER["Scheduler<br/>APScheduler Cron"]
        end
        
        subgraph Data["💾 Data Layer"]
            DB["SQLite + WAL<br/>Singleton Pool"]
            JSON_DB["JSON File Store<br/>Embeddings"]
            DIMS["dimensions.json<br/>190+ Checkpoints (Living)"]
        end
    end

    subgraph External["🌐 External"]
        TAVILY["Tavily API<br/>Web Research"]
    end

    FE_PAGES -->|REST API| ROUTES
    ROUTES --> AUTH
    AUTH --> AI
    ROUTES --> Data
    AI --> GEMINI_PRO & GEMINI_FLASH & GEMINI_EMBED
    EVO_AGENT --> EVO_EXTRACT & EVO_REDUNDANCY & EVO_SCANNER & EVO_SNAPSHOT
    EVO_SCHEDULER --> EVO_AGENT
    EVO_SCANNER --> TAVILY
    EVO_EXTRACT --> GEMINI_PRO
    EVO_REDUNDANCY --> GEMINI_EMBED
    EVO_AGENT --> DIMS

🛠️ Tech Stack

Layer	Technology	Purpose
Frontend	Vite, Vanilla JS, Chart.js	SPA with hash routing, radar charts, force-directed graphs
Backend	Python 3.12, FastAPI, Uvicorn	Async API, 50+ endpoints, 12 routers
AI	Gemini 3.1 Pro, 2.5 Flash, Embeddings	Reasoning, batch eval, vector search, evolution
Evolution	APScheduler, CrossProcessFileLock	Autonomous agent scheduling, safe concurrent writes
Database	SQLite (aiosqlite, WAL mode)	Assessments, analyses, research, evolution history
Research	Tavily API	Automated web research agent, evolution input pipeline
Security	DOMPurify, API Key Auth, slowapi	XSS protection, auth, rate limiting
DevOps	Docker, GitHub Actions, Alembic	Multi-stage build, CI/CD, migrations
Design	Custom CSS (3000+ LOC)	Glassmorphism, dark mode, Inter font

🚀 Quick Start

Prerequisites

API Keys (required):
- Gemini API Key — powers AI features + evolution
- Tavily API Key — powers research agent + evolution scanning
Runtime: Docker (recommended) OR Python 3.12+ & Node.js 18+

Setup

git clone https://github.com/jonasarmbrust/AIStrategyHub.git
cd AIStrategyHub
cp .env.example .env
# ⚠️ Edit .env with your API keys!

Option 1: Docker (Easiest)

docker compose up -d --build

Open http://localhost:8000

Option 2: Local Script (Windows)

start.bat

Builds frontend automatically and starts the backend.

Option 3: Manual Install

# Frontend
cd frontend && npm install && npm run build && cd ..

# Backend
cd backend && pip install -r requirements.txt
python -m uvicorn main:app --port 8000

Open http://localhost:8000

Optional: Enable Authentication

# In .env — set an API key to protect all endpoints
API_AUTH_KEY=your-secret-key-here

Clients must then send X-API-Key: your-secret-key-here header with every request.

Optional: Configure Evolution Agent

# In .env — configure autonomous evolution
EVOLUTION_ENABLED=true
EVOLUTION_SCHEDULE=weekly          # weekly / daily / manual
EVOLUTION_AUTO_INTEGRATE=false     # require human approval
EVOLUTION_MIN_QUALITY=0.7          # minimum quality threshold

📊 The Maturity Framework

The meta-model synthesizes 6 globally recognized frameworks into 7 unified dimensions — and grows autonomously via the Evolution Engine:

Dimension	Weight	Checkpoints	Focus
🎯 Strategy & Leadership	15%	21+	Executive sponsorship, AI-business alignment
🗄️ Data & Infrastructure	15%	14+	Data governance, quality, scalable infra
⚖️ Governance & Compliance	20%	16+	AI risk management, EU AI Act, audit
⚙️ Technology & MLOps	15%	12+	CI/CD for ML, monitoring, deployment
👥 Talent & Culture	10%	12+	AI literacy, cross-functional teams
🛡️ Ethics & Responsible AI	15%	11+	Bias testing, explainability, privacy
🔄 Processes & Scaling	10%	15+	Pilot-to-production, change management

Note

Checkpoint counts shown are the base counts from the original 6 frameworks. The Evolution Agent has expanded these significantly — the living model currently contains 190+ checkpoints and continues to grow.

Source Frameworks

Every checkpoint is fully traceable to its origin:

Framework	Focus Area
NIST AI RMF	Risk management & governance structure
EU AI Act	Regulatory compliance & risk classification
Google AI Adoption Framework	Cloud-native AI scaling
Microsoft Responsible AI MM	RAI practices at scale
OWASP AI Security Matrix	AI-specific security threats
UNESCO AI Readiness	National & organizational readiness

The Evolution Agent continuously discovers and proposes checkpoints from additional sources including arXiv papers, MIT Sloan Management Review, appliedAI, and more.

Maturity Levels

Level	Name	Score Range	Description
1	Initial	0–24%	Ad-hoc, no structured AI approach
2	Developing	25–49%	Early pilots, partial processes
3	Defined	50–69%	Established practices, documented
4	Managed	70–89%	Organization-wide, measured
5	Optimizing	90–100%	Industry-leading, continuous improvement

📁 Project Structure

AIStrategyHub/
├── backend/
│   ├── main.py                    # FastAPI entry point
│   ├── config.py                  # Centralized config & dependencies
│   ├── database.py                # SQLite singleton pool (WAL mode)
│   ├── api/routes/                # 12 API routers (50+ endpoints)
│   │   ├── evolution.py           # 🧬 Evolution Agent (~20 endpoints)
│   │   ├── framework.py           # 🏗️ Framework Builder
│   │   ├── research.py            # 🔬 Research Agent
│   │   ├── advisor.py             # 🧠 AI Advisor
│   │   └── ...                    # analysis, checklist, dashboard, etc.
│   ├── evolution/                  # 🧬 Evolution Engine
│   │   ├── agent.py               # Core Evolution Agent orchestrator
│   │   ├── checkpoint_extractor.py # Gemini-powered extraction
│   │   ├── redundancy_detector.py  # Embedding similarity deduplication
│   │   ├── research_scanner.py     # Tavily deep research scan
│   │   ├── snapshot_manager.py     # Framework versioning & rollback
│   │   ├── scheduler.py           # APScheduler cron scheduling
│   │   └── prompts.py             # LLM prompt templates
│   ├── analyzer/                  # RAG pipeline (embedder, evaluator, parser)
│   ├── knowledge_base/
│   │   └── dimensions.json        # Living Meta-Model (190+ checkpoints)
│   ├── middleware/                 # Auth, rate limiting, error handling
│   ├── models/schemas.py          # Pydantic data contracts
│   ├── research/agent.py          # Tavily research agent
│   ├── migrations/                # Alembic DB migrations
│   └── tests/                     # pytest suite (45 tests)
├── frontend/
│   ├── index.html                 # SPA shell + navigation (15 nav items)
│   └── src/
│       ├── main.js                # Router + API client
│       ├── i18n.js                # Bilingual dictionary (EN/DE)
│       ├── sanitize.js            # DOMPurify XSS protection
│       ├── styles/index.css       # Design system (3000+ LOC)
│       └── pages/                 # 16 page modules
│           ├── evolution.js       # 🧬 Evolution Dashboard (largest page)
│           ├── playbook.js        # 📋 AI Strategy Playbook
│           └── ...                # 14 more pages
├── .github/workflows/ci.yml      # CI/CD pipeline
├── Dockerfile                     # Multi-stage production build
├── docker-compose.yml             # One-command deployment
└── .env.example                   # Environment configuration template

🔒 Security

Feature	Implementation
XSS Protection	DOMPurify sanitization on all LLM-generated content
API Authentication	Optional `X-API-Key` header middleware
Rate Limiting	slowapi with configurable per-endpoint limits
SQL Injection	Parameterized queries throughout
Error Handling	Standardized error responses, no stack traces in production
Input Validation	Pydantic models + file type whitelisting
SSRF Prevention	URL validation with comprehensive security checks

🏗️ Advanced Architecture & Security

AI Strategy Hub implements robust production-grade architecture and security countermeasures:

Concurrency Safety & Lock Isolation: Handles concurrent database and file operations safely. Web routes serialize write operations using an asynchronous database asyncio.Lock, while the Evolution Engine uses a filesystem-level CrossProcessFileLock to prevent database writes from colliding or corrupting the dimensions.json during simultaneous evolution cycles.
Performance Optimization (Batch Embeddings): The RAG analyzer pools text chunks into cohesive batches during document parsing. By issuing batch requests to the Gemini Embedding API rather than sequential calls, it minimizes network latency overhead and increases overall throughput by up to 5x.
Evolution Engine Safety: Framework snapshots are created before every evolution cycle, enabling one-click rollback. The Redundancy Detector uses an SQLite embedding cache to avoid redundant API calls during similarity scans.
Security Hardening & Timing Attack Protections:
- Timing Attack Mitigation: API key verification utilizes constant-time string comparison (secrets.compare_digest) to thwart side-channel analysis aiming to deduce keys.
- DOM XSS Sanitization: User inputs, LLM markdown recommendations, and chatbot outputs are run through a strict DOMPurify pipeline to filter malicious HTML payloads before rendering.
Database Integrity & Cascades: Persisted in SQLite with WAL (Write-Ahead Logging) mode. Enforces strict SQLite Foreign Key constraints (PRAGMA foreign_keys = ON) with cascade rules on deletions to guarantee relational database integrity across analyses, sources, and activities.

📖 API Documentation

The backend automatically generates interactive API documentation:

Swagger UI: http://localhost:8000/docs
ReDoc: http://localhost:8000/redoc

🧪 Testing

cd backend
pytest tests/ -v

Currently 45 tests across 7 test modules covering:

Health & infrastructure endpoints
Scoring engine (8 unit tests with edge cases)
Checklist API (filters, dimensions)
Document analysis (upload, validation, listing)
Security hardening (API key timing attacks, XSS sanitization, SSRF prevention)
Database concurrency safety & constraint enforcement
Evolution engine (safe read/write, concurrent access, redundancy embeddings)

🤝 Contributing

We welcome contributions! Whether it's expanding the maturity model, adding new checkpoints, or improving the codebase — every contribution makes AI governance more accessible.

The most impactful contribution? Expanding dimensions.json with checkpoints from new frameworks. Use the built-in Evolution Agent or Framework Builder to discover and integrate them automatically.

📄 License

MIT License — see LICENSE for details.

⚖️ Disclaimer & Attribution

AI Strategy Hub is an independent, open-source project. It is not affiliated with, endorsed by, or sponsored by any of the organizations whose frameworks are referenced in this tool.

The maturity model synthesizes publicly available concepts from NIST, the European Union, Google Cloud, Microsoft, OWASP, and UNESCO into an independently authored assessment framework. All checkpoint texts are original formulations by the project authors — no content is copied verbatim from any source publication. The Evolution Agent may discover and propose checkpoints from additional academic and industry sources.

Framework names are used solely for attribution and source identification purposes. For authoritative guidance, always refer to the official publications linked in the app's Sources & Attribution section.

Trademark Notice

NIST is a registered trademark of the National Institute of Standards and Technology. Google Cloud is a trademark of Google LLC. Microsoft is a registered trademark of Microsoft Corporation. OWASP is a registered trademark of the OWASP Foundation. All other trademarks are the property of their respective owners.

AI Strategy Hub — Built with Gemini, FastAPI, and a lot of ☕
Assess. Evolve. Optimize.

⭐ Star this repo if you find it useful!

Name		Name	Last commit message	Last commit date
Latest commit History 25 Commits
.github		.github
backend		backend
docs		docs
frontend		frontend
.dockerignore		.dockerignore
.env.example		.env.example
.gitignore		.gitignore
.pre-commit-config.yaml		.pre-commit-config.yaml
CHANGELOG.md		CHANGELOG.md
CODE_OF_CONDUCT.md		CODE_OF_CONDUCT.md
CONTRIBUTING.md		CONTRIBUTING.md
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
SECURITY.md		SECURITY.md
docker-compose.yml		docker-compose.yml
start.bat		start.bat
start.sh		start.sh

Folders and files

Latest commit

History

Repository files navigation

💡 What Is This?

The Problem

The Solution

Who Is This For?

🧬 The Evolution Strategy — Agentic Framework Evolution

How It Works

Evolution Dashboard Features

✨ Key Features

🧬 Evolution Agent NEW

📋 AI Strategy Playbook NEW

🔬 AI Deep Dives NEW

🧠 AI Strategy Advisor

📊 Maturity Assessment

🔍 Document Analyzer (RAG)

🔗 Evidence Chain

⚖️ EU AI Act Compliance

🎮 Gap Simulator

🔬 Research Agent

🏗️ Framework Builder

🕐 Assessment Timeline

🗺️ Strategic Roadmap

🌍 Fully Bilingual (EN/DE)

📸 Screenshots

🏗️ Architecture

🛠️ Tech Stack

🚀 Quick Start

Prerequisites

Setup

Optional: Enable Authentication

Optional: Configure Evolution Agent

📊 The Maturity Framework

Source Frameworks

Maturity Levels

📁 Project Structure

🔒 Security

🏗️ Advanced Architecture & Security

📖 API Documentation

🧪 Testing

🤝 Contributing

📄 License

⚖️ Disclaimer & Attribution

About

Topics

Resources

License

Code of conduct

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases 1

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

🧬 Evolution Agent `NEW`

📋 AI Strategy Playbook `NEW`

🔬 AI Deep Dives `NEW`

Packages