I. The hanout as founding metaphor
In every medina of the Maghreb, there is a man everyone knows. He opens before dawn. He closes after midnight. He knows what every family in his neighborhood consumes, at what frequency, with what margin for negotiation. He manages his inventory by intuition, his cash flow by hand, his customer relationships by memory. He is accountant, logistician, salesperson, purchasing director, and delivery agent. He is the manager of the hanout, that small neighborhood shop which, for centuries, has kept the informal economy of millions of people running.
The hanoutier has no MBA. No board of directors. No HR department. He has something rarer: a total systemic vision of his business, a risk tolerance forged by necessity, and an adaptability that large corporations pay millions to consulting firms to simulate.
I want to talk to you about the Hanoutier 2.0. The one who has replaced memory with databases, intuition with predictive models, on-foot deliveries with autonomous agents. The one for whom artificial intelligence is not an existential threat but the force multiplier he has always been waiting for.
And I want to tell you why juniors are best positioned to become these Hanoutiers of the digital era. Not seniors, not directors, not experts.
II. Ibn Khaldoun understood everything in 1377
In his Muqaddimah, Ibn Khaldoun observes a phenomenon he calls asabiyya, the group cohesion that gives small, tight-knit units their collective strength. Great dynasties, he writes, always end up losing their asabiyya. They grow, bureaucratize, isolate themselves from reality. And it is then that smaller groups, more agile, more cohesive, more motivated, come to overthrow them.
The economic history of the 20th century seemed to refute Ibn Khaldoun. Large companies crushed small ones. Economies of scale, mass standardization, and global distribution networks made it impossible for any individual to compete with General Motors, Procter & Gamble, or Amazon.
But Ibn Khaldoun was right on the substance. He was simply wrong on the timing.
Because what AI has just done is restore asabiyya to the individual.
Schumpeter called it creative destruction, the process by which innovations destroy existing structures to create new ones. What we are living through is not simply a technological wave. It is a fundamental redistribution of productive capacity. For the first time since the industrial revolution, size is no longer a decisive advantage. Speed is. Adaptation is. Precision is.
And these three qualities belong to the Hanoutier.
III. The Hanoutier economics: the math that changes everything
Let us put numbers on the table. Because an essay without numbers is an opinion, not an analysis.
Take an independent AI consultant, a Hanoutier, who charges 800 euros per day. Working 180 billable days per year (a reasonable 70% utilization of a busy schedule), he generates 144,000 euros in revenue.
Now add the AI layer.
That same Hanoutier, with a well-built automation stack, simultaneously manages his business development through an agent that qualifies leads and writes personalized emails, his deliverable production with automatically generated first drafts, his market intelligence and personal branding, his invoicing and accounting through no-code tools, and his client support through a chatbot trained on his own methodologies.
The result? Not that he works less. But that he works differently. Where a traditional consultant spends 40% of his time on administrative and repetitive tasks, the AI Hanoutier spends 10%. The freed 30% goes toward value creation, strategic thinking, and, crucially, scaling his offer.
Here is the equation that changes everything:
Revenue = Day rate × Days billed × Number of simultaneous clients
For a traditional consultant, the number of simultaneous clients is limited by human bandwidth. In practice: 1 to 3 parallel engagements maximum.
For an AI Hanoutier, with agents managing coordination, RAG systems centralizing knowledge, and dashboards monitoring projects in real time, that number can climb to 5, 8, 10 simultaneous engagements at the same quality level.
The difference is not linear. It is exponential.
A senior AI Hanoutier with a mature stack can realistically target 400,000 to 600,000 euros in annual revenue, alone, without hiring anyone. That is the revenue of a 5-to-8-person SME. With margins of 70 to 80% instead of 20 to 30%.
IV. Why juniors are better positioned — not seniors
Here is the thesis that will cause friction, and which I defend without reservation:
Juniors of 2024 are better positioned than seniors of 2010 to become Hanoutiers.
Not because they are smarter. Not because they work harder. But for three structural reasons that seniors cannot easily circumvent.
A senior with 15 years of experience carries 15 years of mental processes, professional reflexes, ingrained ways of working. These habits are assets in a stable world. In a world of rupture, they become liabilities. The junior has none of this cognitive debt. He learns LangChain at the same time he learns to write a project brief. For him, AI is not a new tool added to his practice. It is the foundation of his practice.
The second reason concerns risk. A senior earning 90,000 euros, holding a mortgage, and protecting a reputation cannot afford to fail publicly. The junior has nothing to lose. He can iterate, pivot, experiment. He can launch an imperfect first product, learn from market feedback, and improve it. That is exactly what the most effective Hanoutiers do: they test fast, adjust fast, kill fast what does not work.
The third reason is economic. In 2010, building a consulting practice required a network, an established reputation, and financial resources to sustain the first years without stable income. In 2024, a 23-year-old junior with a MacBook, a Claude Pro subscription, a Vercel account, and six months of skill-building can deliver AI projects of a quality that would have required a team of five people three years earlier.
Ibn Khaldoun would say that the entry barriers protecting the great dynasties have just collapsed.
V. The Hanoutier stack — what it actually looks like
Let us talk technical, because philosophy without execution is literature.
An AI Hanoutier operates with a four-layer stack.
The first is the intelligence layer: Claude for complex reasoning, GPT-4 for creativity, Mistral for sovereign deployments. LangChain or LangGraph for orchestration. A personal RAG system indexing all his knowledge, past missions, frameworks.
The second is the data layer: a structured knowledge base (Notion plus vector database), a lightweight CRM (Airtable or HubSpot), automated intelligence pipelines that read, summarize, and classify continuously.
The third layer handles execution: n8n or Make for workflows, specialized agents for prospecting, client follow-up, and report generation, APIs connected to every tool he uses.
The fourth distributes: a personal website that captures leads (like the one you are reading), content distributed automatically from a single source, a nurturing system that maintains relationships without manual intervention.
The traditional Hanoutier managed all this in his head. The AI Hanoutier manages it in systems. The difference? One is limited by memory and time. The other is limited by the quality of his architecture.
VI. What large companies cannot do
Averroes, in his commentaries on Aristotle, insists on the primacy of direct experience over indirect transmission of knowledge. There is something deeply Aristotelian in the Hanoutier's advantage: he is inside the reality of his client, not above it.
Large consulting firms have a structural problem that AI will not solve for them, because it is not a technological problem. It is an incentives problem.
A major firm charges 2,000 euros per day for a senior and 800 euros for a junior. To maximize revenue, it must maximize billable days. This creates a fundamental tension with efficiency: the more efficient the firm, the less it bills. The Hanoutier, on the other hand, is paid on value created, not time spent. If he delivers in 3 days what used to take 3 weeks, he can charge the value of the 3 weeks and keep the 12 freed days for the next engagement.
Moreover, large companies cannot adapt as fast as the tools evolve. In 2024, new models come out every two months. New frameworks every two weeks. An IT department of 500 people takes 18 months to deploy a new technology. The Hanoutier tests it in 48 hours.
Sun Tzu writes in The Art of War: "Speed is the essence of war." In the economic competition of the AI era, speed of adaptation is the decisive advantage. And the Hanoutier has it.
VII. Equity as argument — not ideology
I want to close on something important, because this essay is not a libertarian pamphlet in disguise.
The rise of the AI Hanoutier is not good news for everyone. It will accelerate the polarization of the labor market between those who master these tools and those who do not. It will create spectacular winners and silent losers.
But it also creates something unprecedented: a path to economic prosperity that is not conditioned by family inheritance, elite school networks, or the first 15 years of a career in a large corporation.
A student in Tunis, Dakar, Lyon, or Montpellier with an internet connection, insatiable curiosity, and the discipline to build his stack can today compete with consultants who have 20 years of experience in top-tier firms.
This is not utopia. It is arithmetic.
The neighborhood Hanoutier always knew that proximity to the client, intimate knowledge of their needs, and the ability to adapt immediately were advantages that big-box stores could not buy, no matter how rich they were.
AI has just given this Hanoutier the tools of the big-box stores. Without removing what makes him strong.
Ibn Khaldoun was right. Dynasties fall. Hanoutiers adapt.
*Saber Dhib is a Data & AI architect based in Paris. He works at the intersection of strategy, engineering, and the delivery of AI systems in production.*
VII bis. Creative destruction is not a metaphor — it is a schedule
Schumpeter described creative destruction in 1942 as "the process of industrial mutation that incessantly revolutionizes the economic structure from within, incessantly destroying the old one, incessantly creating a new one."
What is remarkable in this definition is the word "incessantly." Schumpeter was not describing an event. He was describing a permanent state. The capitalist economy never stabilizes. It destroys and recreates continuously.
What we are living with AI is therefore not an exceptional rupture in history. It is an acceleration of a process that has never stopped. The difference from previous waves, mechanization, electrification, computerization, lies in the speed. And in the target.
Previous waves primarily destroyed physical and repetitive work. They replaced muscles with machines, manual calculations with computers. Complex intellectual jobs seemed safe.
Schumpeter had anticipated this illusion. In Capitalism, Socialism and Democracy, he writes that the classes who believe themselves protected by the complexity of their work are precisely those who will be most destabilized when the next wave arrives. Because they have not developed the resilience that precarity develops.
AI is that wave. And it targets exactly what previous waves had spared: synthesis, analysis, writing, design, consulting. The intellectual fortresses that liberal professions and senior executives believed impregnable.
But creative destruction has two faces. It destroys. It also creates.
What it creates here is unprecedented: the possibility for a single individual to operate with the productive power of an organization. This is not disruption in the startup sense, that of a newcomer attacking a market with a cheaper model. It is a deeper reconfiguration: the disintegration of size advantage as a barrier to entry in high-value-added markets.
Keynes thought technology would liberate humans from work so they could enjoy leisure. He was wrong about human psychology. Most people do not want to work less, they want to work on things that matter. What AI liberates is not leisure time. It is creative capacity. It frees individuals from tasks that consume time without producing thought, leaving them for the tasks that genuinely require human judgment.
For the Hanoutier, creative destruction is not a threat to manage. It is the wind in his sails. Every sector the wave destabilizes is a market that opens. Every large structure that is slow to adapt is a potential client. Every inefficiency exposed by AI is a value opportunity to capture.
Ibn Khaldoun observed that declining dynasties create the conditions for their own replacement by accumulating structures that cost more than they produce. Schumpeter formalized this mechanism. AI has just pressed the accelerator.
The Hanoutier did not cause this destruction. He is simply its best-positioned beneficiary.
VIII. How to become a Hanoutier — organization, training, duration
There is a mistake I see everywhere among those who want to start. They look for the shortcut. They want the perfect stack, the magic prompt, the framework that does everything. They delegate their learning to AI before understanding what they are delegating.
That is the shortest path to mediocrity.
The Hanoutier who succeeds over time is not the one who uses AI best. It is the one who understands enough of what he is doing to know when AI is wrong, when it hallucinates, when its output is mediocre even when it looks convincing. This understanding comes from only one thing: real, repeated, uncomfortable practice.
Mathematics first, not to become a researcher but to develop reliable intuitions. Understanding what a distribution is, what variance means, why a model overfits, how to interpret an ROC curve. These concepts are the foundations on which everything else rests. Without them, you are using tools you do not understand. It is like managing a hanout's inventory without being able to count.
Economics next. The solopreneur's unit economics is the only dashboard that matters: customer acquisition cost, customer lifetime value, net margin per engagement, capacity utilization rate. A Hanoutier who does not master these numbers works without a compass. He may feel busy while he is actually drowning.
Code third. Not necessarily to become a senior developer, but to be able to read an architecture, understand a pipeline, debug an integration that is not working. No-code has its limits exactly where high-value projects begin. Those who cannot code depend entirely on those who can. That is not a position of strength.
The two-speed rule is simple to formulate, hard to hold: learn fast through AI, while deepening slowly through practice without it. Use Claude to understand a concept in 20 minutes, then spend 3 hours implementing it yourself without assistance. AI compresses exposure time. It does not replace experience.
Data on athletic and musical performance all converge on the same conclusion: what separates the excellent from the good is the volume of deliberate practice accumulated over years, not the tools used. AI does not change this equation. It changes the quality and speed of practice. The one who combines both goes further, faster. The one who relies only on AI assistance without practice hits a ceiling he cannot understand why he cannot break through.
In five years, the market will clearly distinguish two categories of Hanoutiers. Those who continued to develop their fundamental skills while mastering the tools, capable of designing complex architectures, diagnosing subtle problems, producing value where AI alone fails. And those who outsourced their thinking to tools without ever developing the intellectual substrate that allows directing them. The former will be rare and very well compensated. The latter will be numerous and interchangeable.
IX. Agents — the Hanoutier's next frontier
I have talked about automation. But I have not yet talked about autonomy. They are different.
An automated workflow executes a predefined sequence. An agent decides. It evaluates the situation, chooses its tools, adapts to unexpected events, and continues toward the objective without human intervention.
For the Hanoutier, agents represent the shift from a model where he supervises each task to a model where he supervises systems that handle the tasks. The difference is fundamental.
Concretely, a mature AI Hanoutier operates with a fleet of specialized agents.
The prospecting agent scans LinkedIn continuously, identifies purchase-intent signals, qualifies prospects against defined criteria, drafts personalized first messages, and submits them for validation before sending. It runs in the background while the Hanoutier works on his engagements.
The sector intelligence agent reads hundreds of sources every day, identifies information relevant to each active client, produces actionable summaries, and flags weak signals. What a junior analyst paid 35,000 euros per year would do, the agent does continuously for a few dozen euros per month.
The document delivery agent produces, from a brief and meeting notes, first drafts of deliverables (architecture reports, delivery plans, risk analyses) that the Hanoutier refines and validates. The ratio of production time to value-added time reverses completely.
The client monitoring agent tracks the progress of each engagement, detects drift risks (delays, satisfaction, scope creep), and alerts proactively with response options.
What was impossible for a single individual, maintaining 8 quality client relationships simultaneously, producing regular content, prospecting continuously, training permanently, becomes not only possible but manageable. Orchestrating this fleet of agents is the distinctive competency of tomorrow's Hanoutier. It is not a technical skill in the traditional sense. It is a management competency: knowing what you want, how to specify it, how to evaluate quality, how to correct drift.
X. The coming disruption — and who will be displaced
Here is what no one says clearly because it is uncomfortable.
Over the next ten years, we will witness a redistribution of economic value unprecedented since the industrial revolution. Not in favor of large platforms, which have already captured their share. In favor of highly competent individuals who will no longer need an organization to amplify their impact.
Supertankers that seemed unassailable will lose market share to fleets of Hanoutiers. Not because these Hanoutiers are better funded. They are not. But because they are faster, more precise, less burdened by organizational debt, and capable of personalizing their offer at a level no large structure can achieve profitably.
Take strategy consulting. A top-tier firm mobilizes a team of 6 people for 4 months to produce an AI transformation diagnostic. Cost to the client: 400,000 euros. A Hanoutier with the right stack, the right market data, and experience with AI architectures can produce equivalent work in 3 weeks for 60,000 euros. The quality difference, if it exists, does not justify a ratio of 1 to 7.
This reasoning applies to management consulting, executive recruitment, professional training, custom software development, compliance auditing, specialized content production. In each of these sectors, well-equipped Hanoutiers will erode margins that large structures considered guaranteed.
The kakistocracy of non-technicians
Something must be named that happened over the past five years and has cost billions of euros to organizations that do not yet know it.
In many large companies and administrations, the hierarchy has been captured by profiles trained to value presentation over substance. Graduates of the top business schools, masters of summarizing what they do not understand, experts in meetings that produce nothing, these profiles thrived in a world where information was scarce and the ability to synthesize it had value.
These are the same profiles who made AI investment decisions. Who hired consultants who resembled them, capable of talking about digital transformation for three hours without ever opening a terminal. Who launched AI initiatives with massive budgets and vague ambitions, led by people who had never trained a model, never deployed an API, never seen a pipeline crash in production.
The result is documented: 80% failure rates on enterprise AI projects according to RAND Corporation. Billions invested for POCs that never reach production. AI training delivered by experts who could not code but knew very well how to invoice.
These profiles created a comprehension debt in their organizations. And that debt will be paid. Not as a fine or sanction. As a growing competitive lag against smaller organizations led by people who genuinely understand what they are building.
The market will realign skills and compensation. It always does. And when it does, it is ruthless.
Those who will be displaced
Let us be precise about who will suffer, because clarity is a form of respect.
Intermediaries without their own added value will disappear first. Those whose role consisted of transmitting information, coordinating flows, producing summaries. AI does all of this better, faster, at zero marginal cost.
Process experts without domain expertise will follow. The project manager who does not understand the project. The consultant who frames without ever delivering. The analyst who produces reports no one reads but everyone orders.
Non-technicians in technical functions will spend the next ten years experiencing a long humiliation. Not a dramatic, visible humiliation. A silent one, measured in decisions increasingly guided by tools they do not control, in budgets allocated according to algorithmic recommendations they do not understand, in recruitment filtered by systems whose criteria they do not know.
And paradoxically, some genuinely good technicians will also be displaced. Those who excelled in a narrow specialty, built on a specific technology, without developing the systemic vision that would allow them to adapt. The developer expert in an obsolete stack. The analyst whose entire value rested on mastery of a tool that AI replaces.
What protects, in this context, is not expertise. It is adaptability anchored in solid fundamentals. The ability to learn fast because the foundations are secure. Understanding of principles, not just tools.
That is why the Hanoutier wins. Not because he has the best tool. Because he understands his neighborhood, his clients, his margins, and he adapts when conditions change. He always has. He will continue.