After 25 years in ICT and a Deputy Director role at Ooredoo, I left to build AI systems full-time. Here's the real story behind the decision.
I spent 25 years in the ICT industry. The last stretch was as Deputy Director at Ooredoo, one of the largest telecom operators in North Africa and the Middle East. Good title. Good salary. Clear career trajectory.
I left to build AI agent systems and run two small businesses.
People ask about this decision regularly. The question is usually some version of “why would you give that up?” The answer is more practical than dramatic.
Telecom is a scale business. Everything operates at thousands or millions of users. The infrastructure decisions affect entire cities. The procurement cycles run in years. The regulatory environment is dense.
Working in this environment for 25 years taught me three things that directly apply to what I do now.
Systems thinking is non-negotiable. A telecom network is a system of systems. Changing one component affects others in ways that are not immediately obvious. This is exactly how multi-agent AI systems work. Change the orchestrator’s routing logic and the effects ripple through every downstream agent. Telecom taught me to think in dependencies before I ever touched an AI system.
Operations at scale are about process, not heroics. In the early years, I thought operational excellence meant having smart people who could solve problems fast. By year fifteen, I understood that it meant having processes that prevented most problems and handled the rest automatically. The smart people designed the processes. They did not firefight every incident.
This principle drove my entire approach to Ayla OS. Thirty cron jobs, automated monitoring, morning briefings, error tracking. The system runs on process, not on me being constantly available.
Business administration matters as much as technology. My Postgraduate Diploma in Business Administration from the University of Wales was not a box-checking exercise. It changed how I evaluate technology decisions. Every technical choice has a business implication: cost, timeline, risk, opportunity cost. Technical leaders who ignore business fundamentals build impressive systems that lose money.
The exact architecture, memory layers, and delegation patterns I use to run 50 agents across two businesses.
Get the AI Agent Blueprint →The decision to leave was not sudden. It was a slow recognition of misalignment.
Telecom is established infrastructure. The problems are real and important, but the pace of change in any large organization is constrained by its size. Proposal to implementation is measured in quarters, sometimes years. Consensus building is necessary and time-consuming.
Meanwhile, AI was moving at a pace where the landscape changed monthly. New models, new capabilities, new architectural patterns. The gap between what was possible and what large organizations could implement was widening.
I found myself spending evenings and weekends building AI projects that moved faster in a week than corporate initiatives moved in a quarter. Not because the corporate work was bad. Because the structures that make large organizations stable also make them slow.
At some point, the side projects stopped being side projects. The time I spent on them was not recreational. I was building real systems, getting real results, and seeing real business potential. The corporate role started feeling like the side project.
Honesty requires acknowledging what I left behind.
Income stability. A senior telecom role pays predictably. Running two small businesses does not. There were months in the first year where revenue from both businesses combined was less than my previous monthly salary. The financial runway had to be planned carefully.
Status. Deputy Director at a multinational telecom operator carries weight in professional contexts. “I run two small companies you have not heard of” does not. This matters less than it felt like it would, but pretending it does not matter at all would be dishonest.
Team. Leading a team of professionals in a large organization is a different experience from managing a small distributed team. The scale of impact is different. The relationships are different. I miss aspects of leading a large team.
Predictability. Corporate life has a rhythm. The fiscal year, the planning cycles, the review periods. You know roughly what the next twelve months look like. Running your own businesses, especially in AI where the market shifts constantly, is fundamentally unpredictable.
Speed. From idea to implementation in days, not months. When I decided to build an automated content pipeline, it was running within a week. In a corporate environment, the requirements gathering alone would have taken longer.
Direct impact. Every decision I make affects the businesses immediately. There is no committee to approve, no consensus to build, no change management process. The trade-off is that there is no safety net either. My bad decisions hit just as quickly.
Relevance. The AI systems I build today use the latest models, patterns, and tools. They are not constrained by enterprise procurement cycles or compatibility requirements with legacy systems. When a better model becomes available, I use it the next day.
Integration of skills. In my corporate role, I used my technical skills and my business skills in separate contexts. In my current work, they are inseparable. Designing an agent architecture is simultaneously a technical and a business decision. Pricing a service engagement requires understanding both what is technically possible and what is commercially viable.
Leaving a prestigious corporate role in Algeria carries specific cultural weight. In North African business culture, stability and corporate prestige are valued highly. The telecom sector, in particular, represents modern, established industry.
Choosing to leave for something unproven, in a field that most people in the region cannot clearly explain, required conversations with family and professional contacts that went beyond the typical “I am changing jobs” discussion.
The business context also matters. Digital Fennec operates in the Algerian market, where AI adoption is early-stage. This means less competition but also less market education. Many prospects need to be convinced that AI is relevant to their operations before you can discuss specific solutions.
Allwebzone operates in the UK market, where AI adoption is more advanced but competition is fierce. Every other consultancy offers “AI solutions.” Standing out requires a clear point of view and demonstrable results.
Running businesses in both markets simultaneously is possible because of the agent systems I built. The operational overhead of managing two companies in two countries is handled largely by automation. Without that, I would need to choose one market.
Yes. Without hesitation, but with more preparation.
If I could advise my past self, three things.
Build the financial runway earlier. I started building my financial buffer six months before leaving. I should have started eighteen months earlier. The stress of uncertain income in the first year was the hardest part of the transition, and it was entirely within my control to mitigate.
Start building the agent system before leaving. I built most of Ayla OS after leaving the corporate role. Building it while still employed, using evenings and weekends, would have given me a more capable system on day one. The ramp-up period would have been shorter.
Invest in the network. My professional network from 25 years in telecom is deep but focused on a specific industry. Building connections in the AI space took time. Starting that networking effort a year earlier would have meant a warmer market when I launched.
The transition from corporate to independent is not about courage or risk tolerance. It is about preparation and timing. Prepare well enough and the risk drops to a manageable level. Time it right and the opportunity is clear.
The newsletter covers the business side of running AI companies, including the numbers that most founders do not share.
Twenty-five years in ICT was not a detour. It was the foundation. Everything I build now stands on what those years taught me about systems, operations, and business. The jump was not away from that experience. It was an application of it in a new context.
Subscribe to The Signal
One issue per week. The biggest AI build of the week, the top five videos worth your time, and an operator's read on what actually matters. For builders who ship, not dreamers who scroll.
No spam. Unsubscribe anytime. Just The Signal.