MMOOZZGG.com

AI department

AI department structure: five roles and who they report to

5 min read1 source

An AI department structure rests on five roles: an AI director, process owners, IT, a lawyer with security, and training. The roles matter because, according to BCG research, about 70% of AI adoption difficulties come from people and processes, 20% from technology and only 10% from algorithms.

On this page · 6 sections

AI director

The AI director answers for the AI solutions map and for the result in money. The director decides which processes to give to AI, in what order, and by which metric to judge the result.

The director turns the knowledge of departments into rules for agents, checks contractors’ estimates and makes sure the rules of different departments do not contradict each other.

Once a week the director reviews agent reports and decides what to fix. Once a month the director brings the top executive the numbers: what is launched, how much was saved, what comes next on the map.

This is not a chief technology officer. A CTO builds IT products; the AI director changes how departments work and answers for the money.

In a large company this is a separate position, also called a chief AI officer. In a small or mid-size company, an outside expert takes the role for a few hours a week.

Process owners

Process owners are the heads of departments: sales, support, finance, HR. Each has a number: the share of leads answered within a minute, the time to close a request, the time to close the month in accounting.

The owner says what AI should do in the process, which cases to pass to people and what the agent must not promise a customer. The owner accepts the agent’s work the same way as an employee’s work.

Process owners do not report to the AI director. They report to the top executive and work with the AI director on AI tasks.

This is where most of the 70% of difficulties that BCG writes about sit. If a department head does not trust the agent, staff redo its work, and there are no savings.

IT department

The IT department connects AI to the company’s systems: CRM, 1C accounting, phone lines, messaging apps. There is no need to change these systems for AI; agents work with what already exists.

The IT department gives each agent a separate account with rights only to its own process. This shows what the agent did and what a person did, and access can be closed at any moment.

The IT department does not have to decide how sales work will change, and it cannot. It answers for keeping everything running and for the security of systems, while the department sets the task.

One more IT task is the change log. Each edit to an agent’s rules is recorded and can be rolled back if the metric drops after it.

Lawyer and security

The lawyer decides where customer personal data is processed and checks the contract with the AI supplier for compliance with 152-FZ, Russia’s personal data law. Until this check, the agent gets no access to the customer base.

The lawyer also decides what the agent may say on behalf of the company: prices, terms, discounts. The company answers for the agent’s promise to a customer, so the limits are written down in advance.

The security team keeps a list: which information may go to outside AI models, which only to models inside the company, and which to nobody. Every employee gets this list.

Security also checks access: the agent sees only what its process needs. The list of each agent’s rights is kept by the IT department and reviewed for each new process.

Training

The training role teaches staff to work together with AI: what the agent does, what stays with a person, how to accept its work. Training uses the company’s own processes, not textbook tasks.

Department heads learn to choose a process for AI and to judge the result by a number. New employees start their work with agents already in place.

Training is not a one-time event. Models, prices and laws change every month, so the training role tells departments what has changed and how it affects their work.

In a small company the AI director combines this role. In a large one it is a separate person or a training service that works by the department’s plan.

Who the department reports to

The AI department reports to the top executive: the CEO or the owner. AI changes how sales, accounting and production work, and that is a decision for the top executive.

Reporting to the CIO narrows the task to technology, that is, to the 20% and 10% of difficulties in the BCG research. People and processes are left without an owner.

The top executive sets the goal and approves the AI solutions map. The AI director brings the plan, price and term, process owners give numbers for their departments, and IT and the lawyer say what is allowed and what is not.

The report goes to the top executive regularly and in money. Which indicators to include is covered on the page on department metrics.

Sources and fact-checking

Checked by the editors on 13 October 2026.

  1. BCG: 74% of companies get no tangible value from AI, 24 October 2024 (bcg.com)

Found a mistake? Tell us

Publications

Contact us