We tend to assume that the AI era will require even more programmers. Perhaps it will. But some of the most valuable professionals of the next decade may look very different: people who understand technology well enough to direct it — and understand a real industry deeply enough to know where technology actually belongs.
Imagine Yerevan in 2030.
A major bank is hiring someone to lead an AI project.
There are two candidates.
The first is an exceptional AI engineer.
They understand models, architectures, APIs, agents and machine learning. Ask almost any technical question about modern AI, and they can answer it.
The second candidate understands AI too.
Perhaps not quite as deeply from a purely technical perspective.
But they have spent ten years in banking.
They understand lending.
They know why certain operations cannot simply be automated.
They understand compliance.
They know where employees lose time.
They have watched promising technology projects fail after implementation.
And they can look at a banking process and say:
“AI could fundamentally improve this part. But putting AI over there would solve almost nothing.”
Who does the bank hire?
A few years ago, the answer might have seemed obvious.
Today, it doesn't.
And within a few years, this question could become one of the defining questions of the labor market.
We Have Spent Too Long Dividing the World Into “Technical” and “Non-Technical”
There are programmers.
And there is business.
There are engineers.
And there are financial professionals.
There is IT.
And there is medicine.
There are developers.
And there are lawyers.
Organizations themselves were built around this separation.
IT sits over here.
The business sits over there.
One side explains what needs to be done.
The other figures out how to build it.
That model worked for decades.
AI is beginning to blur the boundary.
Because technology is simultaneously becoming easier to use — and harder to apply intelligently.
That sounds contradictory.
It isn't.
Using AI Is Getting Easier. Knowing Where to Use It Is Getting Harder.
You no longer need to be a programmer to interact with powerful technology.
You can describe a task in natural language.
Build a prototype.
Analyze data.
Generate code.
Automate part of a workflow.
Conduct research.
Create an agent.
The technological barrier is falling.
But another barrier is becoming more important:
What should we actually automate?
How do we distinguish an AI use case that creates measurable value from one that merely adds an impressive interface?
Which decisions can be delegated?
Where is an error tolerable?
And where could one incorrect recommendation cost a company millions?
You cannot find those answers in the documentation of an AI platform.
You need to understand the profession itself.
Imagine a Doctor Who Understands AI
They do not necessarily need to know how to train a neural network from scratch.
But they understand:
which medical data matter;
which mistakes are critical;
where AI can assist a physician;
where human judgment is essential;
how a real clinical workflow operates;
which outputs can be trusted and which require verification.
Now imagine an outstanding AI engineer who has never worked in healthcare.
Who is more likely to discover the most valuable application?
The best answer may be:
both — if they can truly understand each other.
But there is a third person who becomes especially interesting.
Someone who understands both worlds.
This May Become a New Category of Professional
Not simply a programmer.
Not simply a banker.
Not simply a doctor.
Not simply a lawyer.
But:
AI + Banking
AI + Cybersecurity
AI + Medicine
AI + Law
AI + Manufacturing
AI + Telecommunications
AI + Logistics
AI + Agriculture
AI + Finance
This does not mean every accountant needs to become a data scientist.
Nor does every software engineer need a medical degree.
The point is different.
The intersection between technological competence and domain expertise may become one of the most valuable places in the labor market.
Why This Matters Especially for Armenia
Armenia has spent years developing substantial technological human capital. In 2026, that effort is broadening further: national initiatives are expanding digital and AI literacy, while education programs increasingly emphasize AI skills, critical thinking and problem-solving. UNESCO's recent assessment of Armenia also highlights both the country's technology-sector growth and the need to keep ICT education aligned with labor-market requirements.
But the next challenge may be different.
It is not simply about producing more people capable of building technology.
It is about producing more people capable of connecting technology to the real economy.
Because AI by itself does not modernize a bank.
It does not transform manufacturing.
It does not improve healthcare.
It does not optimize logistics.
It does not reinvent agriculture.
That happens only when someone understands both:
what the technology can do
and
how the industry actually works.
AI Creates a Strange Paradox
The more powerful technology becomes, the more valuable context becomes.
AI can analyze a thousand pages in minutes.
But who determines which pages actually matter?
AI can discover patterns in data.
But who determines whether those patterns make business sense?
AI can recommend a solution.
But who knows whether that solution can survive contact with a real organization?
AI can write code.
But who decides which problem is worth solving with that code in the first place?
This leads to an important shift.
Understanding the problem may become as valuable as understanding the tool.
The Most Valuable Person in the Room May Not Be the Best Programmer
It may be the person sitting between two worlds.
In a meeting with engineers, they understand the architecture.
With the CEO, they understand the economics.
With the CFO, they can discuss ROI.
With cybersecurity, they understand risk.
With employees, they understand the actual workflow.
With the AI team, they understand data and models.
Most importantly, they can translate.
Not from Armenian into English.
From the language of business into the language of technology — and back again.
These people have always existed.
AI may dramatically increase their value.
Because the Next AI Bottleneck May Not Be Technology
There will be more models.
More compute.
More platforms.
More APIs.
More copilots.
More agents.
More ready-made AI capabilities.
Armenia itself is investing in the other side of this equation: access to high-performance computing, AI infrastructure and broader AI capacity is becoming part of the country's technology agenda.
That creates an interesting possibility.
In the future, obtaining access to powerful AI may become easier than finding someone who knows:
what to do with it.
And that changes the competitive equation.
If ten competitors can access similar technology, the technology itself becomes less differentiating.
The advantage moves elsewhere:
to the ability to apply it better.
Two Banks Can Buy the Same AI
One deploys a chatbot.
The second goes deeper.
It identifies dozens of processes where employees repeatedly perform intellectual tasks.
It determines which data are required.
It redesigns workflows.
It decides where human judgment must remain.
It changes KPIs.
It trains employees.
It integrates AI into existing systems.
One year later, both organizations can truthfully say:
“We use AI.”
But the economic outcomes may be completely different.
Why?
They had access to similar technology.
The difference was not the AI.
The difference was the organization's ability to understand where to put it.
And This Changes the Meaning of Technology Education
Perhaps we should stop asking young people only:
“Do you want to learn programming?”
And start asking:
“What problem do you want to learn to solve with technology?”
For one person, the answer will be cybersecurity.
For another, medicine.
Finance.
Energy.
Architecture.
Manufacturing.
Biology.
Logistics.
Public administration.
The combination of those disciplines could produce an entirely different generation of professionals.
And Armenia is already moving toward broader AI literacy: current initiatives include nationwide digital-skills programs and plans to expand AI education across the school system.
The University of the Future May Look Different Too
Imagine a medical student.
Alongside medicine, they learn:
data literacy;
AI fundamentals;
AI ethics;
model evaluation;
data privacy.
Now imagine a finance student who also understands:
automation;
APIs;
data analytics;
AI agents.
And an engineer who studies not only technology, but the actual business processes of the industry they intend to transform.
The objective is not to create people who know a little bit about everything.
It is to create professionals with two forms of depth.
That distinction matters.
Because Superficial Knowledge of Two Fields Is Not Enough
It would be easy to misunderstand this argument.
The future does not necessarily belong to generalists who know “a little AI, a little finance, a little marketing.”
Quite the opposite.
The valuable combinations may be:
deep domain expertise + meaningful technological competence
or
deep technological expertise + serious understanding of a specific industry.
That is much harder to develop.
Which is precisely why it may become so valuable.
Hiring Will Change Too
A job description may eventually require more than:
“Five years of banking experience.”
Or:
“Three years of AI experience.”
Companies will increasingly search for combinations.
Someone who understands banking risk and data models.
A cybersecurity professional who understands AI agents.
A lawyer capable of designing AI-assisted document workflows.
A sales professional who can build an AI-driven operating process.
A manufacturing engineer who understands computer vision.
New professions will emerge at the intersections.
But the Biggest Change Will Not Be for Employees. It Will Be for Companies.
An organization can buy technology.
It cannot simply purchase the organizational ability to use technology well.
That capability has to exist inside the business.
Which suggests a different question for CEOs.
Not:
“How many AI specialists do we employ?”
But:
“How many people in our organization understand both our business and what AI can actually do?”
If the answer is two people out of five hundred, that may be the real bottleneck in the company's AI transformation.
Not software.
Not GPUs.
Not even budget.
People capable of connecting all of those things to a real problem.
For Armenia, This Could Become a Major Opportunity
Armenia already has a technology ecosystem.
Engineers.
Entrepreneurs.
International experience.
Growing AI infrastructure.
The next stage could be more interesting still.
Not only Armenian software engineers.
But Armenian:
AI-finance specialists;
AI-security experts;
AI-biotech researchers;
AI-manufacturing engineers;
AI-education specialists;
AI-agriculture experts;
AI-legal professionals.
At that point, technological capability begins spreading through the wider economy.
IT stops existing as an isolated island.
It becomes part of every profession.
Perhaps That Is What Real Digital Transformation Looks Like
Not when every company buys new software.
Not when every website gets an AI chatbot.
Not when the number of IT companies increases again.
But when doctors, bankers, engineers, lawyers, entrepreneurs and public-sector professionals begin to understand:
what technology can actually do inside their own profession.
Then AI stops being merely a product of the technology industry.
It becomes part of the infrastructure of the economy.
Final Thought
For much of the last two decades, the technology race has also been a race for engineers.
Who can attract the best developers?
Who can educate more programmers?
Who can create the strongest technology ecosystem?
That race is not disappearing.
But another race is beginning alongside it.
The race for people who can connect knowledge.
People who understand not only how AI works, but why a particular industry needs it.
Because technology is gradually becoming available to everyone.
Deep understanding of how to convert that technology into real value is not.
And that is why the most valuable professional of the AI era may not be the person who knows artificial intelligence better than everyone else.
It may be the person who can look at a real problem inside their industry and say:
“I know exactly where AI belongs here. And I know where it doesn't belong at all.”
Perhaps those are the people Armenia's next technology race will really be about.