Protection from neural networks: where the money of Kazakhstan's IT sector lies in the era of AI agents

Askar Aytov, city — Almaty, DevRel and founder of the Telegram channel Devs.kz

Incomes of American IT specialists have begun to decline. The technology sector in the US has already recorded a drop of 3.4% as of March 13, 2026. Historically, such trends reach Central Asia with a lag of about a year or a year and a half.

The main question for developers today is how to preserve and increase incomes when autonomous AI agents enter the market.

To understand the scale of possible changes, it is worth looking at how forces are currently distributed in the Kazakh IT sector. We are talking about tens of thousands of specialists:

— outsourcing: about 3,000 people. This includes companies servicing government contracts, as well as well-known local players, and branches of international giants like EPAM.

— FinTech and banking: the seven largest banks hold around 2,800–3,000 people in their digital, core IT, and data science departments.

— telecommunications: three to five key market players accumulate no less than 1,000 developers.

— product companies and startups: about 3,500 specialists. There are approximately 300 startups and 30–50 large product companies at the level of ONAY or Documentolog operating in the country. On average, such teams employ more than 10 people.

— hidden reserve and relocators: this is a large segment that is difficult to estimate accurately. This includes the Kazakh group of companies Freedom Holding, relocated corporations Yandex, Ozon Tech, and T-Bank, as well as specialists working for Western markets.

Additionally, about 10,000 engineers and IT officers serve in departmental structures, and the academic IT sector — teachers and research staff — numbers no less than 500 people.

Paradigm shift: from chatbots to AI agents

The market will no longer be the same.

One Almaty-based outsourcing company, working in the US market, has already fully transitioned its employees to AI-assisted development. Business analysts, testers, and programmers work in tandem with autonomous agents.

Globally, the industry is moving toward a Service as Software model. Salesforce, with its Agentforce product, is implementing payment for specific results rather than for providing information.

In the DevOps community, AI technical support engineers are already appearing who autonomously resolve incidents, reduce MTTR — system recovery time — and prevent unnecessary ticket escalation.

On the local Kazakh market, agents in full production are still rare. One can highlight only Bids.do.

Attempts to implement Spec Driven Development on pet projects show that manual work in pairs with ordinary chatbots is gradually becoming a thing of the past. Maintaining code quality requires transitioning to new orchestration tools such as Google Antigravity, Claude Code, or Cowork. The technological gap will only grow.

Enterprise systems — how they protect themselves from vibe coders

If code becomes a commodity that neural networks can produce quickly and cheaply, where should developers find a sustainable niche?

The answer lies in the B2B segment and a deep understanding of business processes.

A good example is Documentolog, a large Kazakhstani document management system from Astana. Artificial intelligence could not displace this product from the market. On the contrary, the company itself began integrating AI agent generators into the system.

What protects enterprise development is what is not present in neural network training databases — an understanding of complex and often undocumented business processes.

Large enterprises in the real sector, for example in water supply or heavy industry, have complex integration architectures. Artificial intelligence can write code, but it does not know how a specific enterprise operates in practice and at the level of bureaucracy.

Reports from corporations at the Salesforce level confirm this thesis: profits in the enterprise segment remain stable even amid technological turbulence.

Where to go next

Money is increasingly concentrating in the B2B sector.

Geographically and strategically, this capital traditionally revolves around Astana. It is quite likely that in the near future we will see a surge of new, albeit short-term, opportunities around the under-construction G4 City in Alatau, where digital infrastructure will need to be built from scratch.

The future of Kazakh IT is not for those who simply write code quickly, but for those who understand business logic and know how to delegate routine tasks to AI agents.