AI Operations Lab
for Orange Romania
We train your people to build AI systems and architect the patterns they build on. In 18 months, you don't need us. That's the point.
The Reality
Orange has the customers.
But AI isn't working yet.
Contract Renewals, 5-Day Wait
Support Emails: Manual Triage
Romanian NLP: Underdeveloped
Shadow AI: No Governance
€600M AI Value Target: Clock Ticking
No WhatsApp Presence
Seen enough? Start with €30K-€50K.
No procurement committee. No RFP. One conversation.
What We Offer
We transfer AI capability.
You build it. You own it. You run it.
IBM and Accenture deliver products, and a renewal invoice. We deliver an architectural capability that lives inside Orange permanently. Your people build the agents. The methodology is ours. The outcome is yours.
The core principle
The agents handle the volume. Your people handle what requires judgment.
These AI agents are not here to replace Orange employees. They exist to remove the high-volume, repetitive grunt work, so your team spends their time on decisions, relationships, and complex cases. The work only humans should be doing.
Train
8-12 Orange employees learn to build AI agents. Not a course, a build sprint. They leave with skills, not certificates.
Build
Each team builds a working agent for a real Orange problem. Contract renewals. Email triage. The agent takes the volume. The employee takes the credit.
Architect
What your team builds reveals your actual integration landscape, turning 10 weeks of hands-on work into an 18-month AI roadmap grounded in evidence.
The Value for Orange
What this fixes.
Specifically. For you.
You already know which workflows are painful. Phase 1 picks two or three and proves them.
Sponsor-led selection from a menu of five (support and service triage, retention and renewal copilot, internal knowledge assistant, network triage, procurement). Voice surfaces like Djia are one option, not the assumption. Whatever Orange picks becomes the evidence base for Phase 2.
Paris set the target. Bucharest needs to deliver.
Phase 1 produces 3-4 working agents and an architecture blueprint. Bucharest becomes the Group's AI reference implementation, political capital for your leadership.
15 internal experiments. No coordination.
12 trained staff who speak the same AI architecture language. A governance framework that coordinates what's already happening internally, and scales it.
Your stack already belongs to someone else.
Every agent your team builds belongs to Orange. The architecture blueprint lives on InfoAcademy. Zero APEX dependency after Phase 1 if you choose not to continue.
Track Record
We already built this.
For ourselves and for clients.
GTM Agentic Platform
Co-Founder, Cognition Architecture
Co-founding a plug-and-play GTM platform (€33B TAM) with an established EU procurement founder. Enterprises connect their vendor and customer data. The cognition layer handles discovery, qualification, matching, and onboarding autonomously. Full vendor lifecycle outsourced to AI, not to a consulting firm. Built on €3M+ of existing R&D.
In DevelopmentYour Team
Enterprise engineering.
AI-native delivery.
IBM, Sun Microsystems, and Lloyds pedigree. 60+ years enterprise delivery. Cloud, security, and SCADA engineering. Delivered through InfoAcademy, your existing vendor.
Strategic Leadership
Nico Fratila
Founder & CEO of APEX OS, the sovereign AI operating system running 24/7 on Azure with 400+ self-evolving skills, persistent agent memory, multi-model orchestration, self-healing infrastructure, and a knowledge distillation engine. Also runs InfoAcademy, Orange Romania's existing training vendor. Background spans 5 years of network engineering at Lloyds Banking Group (cloud migrations and operational transformation), business analysis and project & programme management across FinTech and RegTech, and scalable platforms built from scratch. Chief Customer Office at CUBE Global. Helps founders and businesses move from concept to execution with battle-tested frameworks, AI automation, and hands-on technical expertise. Architect, overseer, cognitive-intelligence layer, deployer, shipper.
Mru Patel
40 years at the intersection of innovative, disruptive enterprise technology and global commercial strategy. Led multi-billion-pound programmes at IBM, Siemens, and Sun Microsystems, including pioneering several vertical-sector solutions including telcos, open-source platforms, and securing deals valued at over £3B across Europe, Africa, and Asia. Government and Fortune 500 advisor since 1987. Established the first independent Property Investment Fund in Eastern Europe in 2004. Founder and CEO of several startups in FinTech, HealthTech, Marketing & Comms. Brings the strategic and commercial architecture that turns a strong technical proposal into a Group-level decision.
Execution Team
Bogdan Toporan
20+ years leading engineering teams at Telenav, Micro Focus (HP), and Emerson, where he patented a SCADA communication system for the Oil & Gas sector. Bridges traditional IT and the AI revolution through two consulting brands: Arandi Software takes brittle monoliths to resilient, scalable cloud systems (Microservices, Kubernetes, AWS); Digitalize Today designs and deploys agentic AI, workflow automation, and intelligent document processing that actually works in production, saving hundreds of hours of manual labor.
Hardik Nakum
Over a decade across global financial institutions at the intersection of AI, cloud, and cybersecurity, driving enterprise transformation with a strong focus on resilience, governance, and long-term impact. Has led large-scale cloud and transformation initiatives across Azure, AWS, and GCP, supported multi-cloud re-architecture programmes, and helped leadership teams adopt innovation in a way that is practical, responsible, and aligned to business outcomes. Builds the next generation of enterprise systems where humans and AI work together.
Local & International Partners
Nautilus OT
OT cybersecurity · IndustrialOperational Technology cybersecurity specialists. 'Securing Your OT Future.' Tracks OT/ICS incidents across shipping, oil & gas, and telecoms; publishes the OT Cyber Insights biweekly update.
The Engagement
Three phases.
18 months to AI independence.
Start at Phase 1. Scale when ready. Stop any time, you keep everything.
AI Operations Lab
Funded from L&D budget · Below VP discretionary spend · No procurement committee
10-Week Program
Foundations
Agent patterns, orchestration, deterministic vs probabilistic pipelines, governance basics
Build Sprint
Each team picks a real Orange use case (network triage, knowledge base, procurement analysis). Orange chooses the workflow. We guide the build.
Architecture Review
APEX reviews what your teams built. Synthesises patterns. Maps integration opportunities against Orange's live systems.
Roadmap Delivery
Living blueprint on InfoAcademy, not a PDF. 18-month roadmap grounded in what your people built, not external assumptions.
What You Get
3-4 Working AI Agents
Built by YOUR team. Owned by Orange.
Living Architecture Blueprint
On InfoAcademy, not a PDF. Evolves with your stack.
12 Trained Employees
Your internal AI champions. Ready to scale Phase 2.
18-Month AI Roadmap
Grounded in your builds, not consultant assumptions.
Governance Foundation
Standards, audit trails, shadow AI visibility.
Capability Assessment
Evidence for your Group AI mandate reporting.
Operationalise & Integrate
From Phase 1 winners to production
Enterprise AI Operating Model
Governed, scalable, durable. Owned by Orange.
Straight Answers
Questions you should be asking.
Why not hire an AI engineer internally for €5K/month?
You're a small company. Can you handle Orange scale?
What if we take the blueprint and give it to Accenture?
Orange Group has AI teams in Paris. Why go external?
€30K-€50K seems cheap for 'architecture' but expensive for 'training'.
Ready to build?
10 weeks. €30K-€50K based on scope. Working AI agents built by your team.
If the agents don't deliver value, you still have 12 trained people.
InfoAcademy × APEX OS: Bucharest, Romania · London, UK
nico.apex.infoacademy.uk