What AI transformation is
It is a rebuild of how a company works around AI: which processes to give to AI agentsAI agent. By definition, an AI agent is a neural network that does not just answer but carries out a task step by step on its own: it looks for data, uses programs and checks the result., who owns them, how people’s roles change and how to measure the result in money.
It differs from introducing one tool. Any department can buy an AI subscription in a day. Transformation changes the process from order to payment, and that process runs through sales, the warehouse and accounting.
Example: in distribution, an order arrives in a messenger, a manager moves it to the CRM, the warehouse checks stock in 1C, a Russian accounting system, and accounting issues an invoice. After transformation, an AI agent reads the order and prepares the stock check and the invoice itself; the manager confirms disputed orders.
Transformation has three signs. The CEO makes the decision, not the IT department. Each solution has a metric in money. The company gets people and rules that are responsible for AI after launch.
Large companies take this path with partners. According to a survey by a major Russian tech company and the consulting firm Yakov and Partners among 150 CTOs of large companies in 16 industries, 78% use ready-made AI solutions from outside partners, and 71% fine-tunemodel fine-tuning. Model fine-tuning is the adjustment of an already trained model on your own examples, so that it performs a specific task more accurately or writes in the required style. ready models instead of training their own from scratch.
How it differs from digital transformation
Digital transformation moves processes into systems: paper into 1C, calls into the CRM, reports into dashboards. People still do the work, only in software. AI transformationAI transformation. By definition, AI transformation is the rebuilding of a whole company’s work around artificial intelligence: processes, roles, rules and the way decisions are made change, not only the set of programs. goes further: AI does part of the work inside these systems.
Digital transformation must come before AI: if orders live in personal messengers and stock in a notebook, there is nothing to connect AI to. You cannot run both on one plan: they have different owners and different metrics. The CEO leads AI transformation, while the IT department connects AI to the systems and manages access.
A detailed comparison of who is responsible and how to measure the result is in the article on the difference between digital and AI transformation.
The program: from review to scale
The program goes in steps, and each next step starts from the numbers of the previous one. Each step ends with a decision by the CEO: go on, rework or stop.
The first step is AI Due Diligence, a review of processes, data, systems and AI solutions already bought, across all departments. The second step is the AI solutions map: where AI returns money, with which solution, what it costs and in what order to introduce it.
Next come pilots: one AI agent on one process with a metric in the contract. Then the AI departmentAI agent team. An AI agent team is a group of AI agents that run a company process together: each agent plays its own role, and a person sets the rules and checks the important decisions.: an AI director, process owners, the IT department, a lawyer and security. Then scale to several departments and training staff to work with AI on their own tasks.
What happens at each step and what the company gets at its end is covered in the article on the stages of AI transformation.
What the company gets
An answer to where AI returns money, with a calculation in money for each solution. Working solutions with a metric in the contract: what it was, what it became, how much was saved. An owner for each solution, data rules under 152-FZ, the Russian personal data law, and control of vendors and budget.
An outside partner raises the chance of reaching a result. In the MIT NANDA 2025 report, projects with an outside partner reached real use in about 67% of cases, and in-house builds in 33%. This is survey data, not an exact measurement, but the gap is twofold.
For large business, the program covers the whole company: a review of all departments, the AI department design and a plan for the year.
Experience: 120 engineers became an AI team
The program is led by Mikhail Mozg, an AI directorChief AI officer. A chief AI officer is the executive who is responsible for artificial intelligence across the whole company: chooses what to implement and controls contractors, the budget and the risks. and AI systems architect with 20 years in software development. At Devblock (2020–2026) he was CTO and moved an IT company of 120 people to development with AI agents.
The same approach works for clients: start with processes where work repeats, give AI agents access to working systems, write down the rules, test on a pilot, then scale. How it worked is covered in the story of the AI transformation of a company of 120 engineers.

