Fortis Case Study: Private AI for $500B+ Client Wealth
Last Updated: January 8, 2026
R6S built "Alfred"—a private AI system for Fortis, the global leader in secure private travel serving clients with a minimum net worth of $600 million. Fortis coordinates over 25,000 trips across 141 countries annually for clients collectively worth over $500 billion. Alfred runs entirely on Fortis hardware with zero cloud dependencies, delivering 6+ months of work in 3 months.
TL;DR
- • R6S built "Alfred" private AI for Fortis (clients worth $500B+)
- • Delivered 6+ months of work in 3 months, ahead of schedule
- • 100% on-premise, zero cloud dependencies
- • CEO testimonial: "I trust Ash with no reservations"
The Client: Fortis
Fortis is the global leader in secure private travel for ultra-high-net-worth individuals. They coordinate over 25,000 private, secure trips across 141 countries every year.
Their clientele is collectively worth over half a trillion dollars. The minimum net worth of a Fortis client is $600 million. They serve Forbes 400 billionaires with secure chauffeured transportation, executive protection, and concierge services.
25,000+
Trips per Year
141
Countries
$600M+
Min Client Net Worth
$500B+
Client Wealth Served
Their entire business is built on trust and discretion. When you're responsible for the safety and security of billionaires and their families, there is no margin for error. That's the standard Fortis operates at every day.
The Challenge
Fortis needed AI capabilities but couldn't risk their clients' sensitive data touching third-party servers. No cloud AI. No data leaving their control. They needed a system they could own outright - one that would give them AI advantages while maintaining the ironclad security their clients expect.
They needed someone they could trust with their most sensitive operational data. Someone who understood that privacy isn't a feature - it's the foundation.
The Solution
R6S built "Alfred" - a private AI system that runs entirely on Fortis's own hardware. No cloud dependencies. No third-party access. Complete ownership.
Private AI Brain
Custom AI trained on Fortis's operations, running locally on their machines. No data ever leaves their control.
Agentic Automation
AI agents that can take action - searching, drafting, processing - with human-in-the-loop controls for safety.
Encrypted Backup System
256-bit encrypted backups that Fortis owns completely. Even on cloud storage, the data remains unreadable to anyone else.
Team Independence
Full training for the internal team. Fortis can maintain, customize, and expand the system without ongoing dependency.
The R6S Process: How Alfred Was Built
Every R6S engagement follows a proven methodology. Here's how we built Alfred for Fortis:
-
1
Discovery Phase
2 weeks understanding Fortis operations, workflows, and security requirements. Mapped every department's AI needs.
-
2
Architecture Design
Designed multi-model local AI system with 120B parameter main model, 20B mini model, and 27B vision model. Zero cloud dependencies.
-
3
Development & Integration
Built custom department-specific AI personas (Security Alfred, Accounting Alfred, etc.) with RAG knowledge bases and GD2 database integration.
-
4
Team Training
Full independence training for Fortis team. They can maintain, customize, and expand Alfred without ongoing R6S dependency.
-
5
Delivery
Completed ahead of 3-4 month estimate. Added bonus encrypted backup system beyond original scope.
+
Technical Deep Dive
Click to expand
For those who want to understand the technical depth of what was built:
Multi-Model AI Architecture
Alfred runs multiple AI models locally on Fortis hardware, each optimized for different tasks:
- • 120B parameter main model - Primary advisor for complex reasoning and strategy
- • 20B parameter mini model - Faster responses for day-to-day queries
- • 27B parameter vision model (Gamma) - Processes images and visual documents
- • Challenger model - Verifies and stress-tests responses for accuracy
Department-Specific AI Personas
Custom system prompts create specialized AI assistants for each team:
Knowledge Base & RAG System
- • Upload files, folders, or raw text to create searchable knowledge bases
- • Retrieval Augmented Generation (RAG) with intelligent chunking and overlap
- • Dual-model ranking and re-ranking for accuracy verification
- • GD2 database integration via custom-built query tool
Voice & Meeting Intelligence
- • Record meetings and transcribe via Whisper (speech-to-text)
- • AI-enhanced notes that extract dates, attendees, and action items
- • Text-to-speech for hands-free AI responses
Security Team Workflow
- • Pre-built prompt templates (Step 0-10) for standardized AAR reports
- • Web search integration (Perplexity API) for real-time intelligence
- • Location-based security assessments with current events context
Enterprise Controls
- • User management with role-based access
- • Temperature controls for creativity vs. accuracy per use case
- • Custom functions (e.g., brand compliance filters)
- • Evaluation tracking via thumbs up/down feedback
All of this runs on Fortis's own hardware. No cloud. No third-party access. Complete ownership.
The Results
6+ Months
of work delivered in 3 months
10 to 4
deployment steps automated
100%
data ownership retained
The project scope was estimated at 3-4 months. It was delivered ahead of schedule with additional capabilities beyond the original scope - including a backup system that wasn't part of the initial agreement but was built because it was the right thing to do.
In Nathan's Words
Nathan Foy, Founder & CEO of Fortis, wrote the following recommendation:
FORTIS
EXPECT MORE FROM TRAVEL
To whom it may concern:
I am writing to recommend the work of Ash Sobhe and his company, R6S. Ash did tremendous work for my company and I'd recommend him to anyone needing help on data security, artificial intelligence applications, or consulting in those areas.
Ash took the time to get to know our team and our needs in a way that considered the whole picture. He then provided a detailed roadmap with clear deliverables and metrics that helped us know what to expect. Having limited technical expertise, this was very helpful for me to know what to expect and how to explain to the team how to make the most of Ash's work.
Ash completed the assignment ahead of pace and did much more than expected. He communicated with the team and provided timely updates. Trust is a very key component to have on a person doing this kind of work with your most sensitive data. I trust Ash for his integrity and performance with no reservations.
I couldn't recommend Ash and R6S more highly for your consideration.
Sincerely,
Nathan Foy
Nathan Foy
CEO
Fortis Corp.
fortis.co | US East | US West | Hong Kong | India
"Trust is a very key component to have on a person doing this kind of work with your most sensitive data. I trust Ash for his integrity and performance with no reservations."
Nathan Foy
CEO, Fortis Corp.
Learn More
Ready to build something similar?
Private AI. Complete ownership. Real results.
Start a Conversation