Who Should Be Responsible for Artificial Intelligence Within the Company? Profile, Responsibilities, and 3 Models by Size
By Pedro Sonda, dooop
Responsibility for AI should rest with a person appointed by the executive team, with a formal mandate, business vision, cross-departmental access, and dedicated time for the role, supported by IT, legal, and risk teams. It should not be "everyone's responsibility" nor exclusively IT's. This person has five key responsibilities: leading the agenda, managing the portfolio of initiatives, ensuring governance, coordinating training, and reporting to the executive team. The structure depends on company size: a responsible person combining the role with a small core team, a dedicated responsible person with a committee, or an AI office.
Updated October 2026.
What the Data Shows
| Data | Source |
|---|---|
| In a Gartner poll, 54% of organizations had a head or leader of AI, but 88% of those leaders did not hold the title Chief AI Officer; only 25% of respondents indicated a clear role responsible for AI | Gartner, 2024 |
| 91% of organizations with high AI maturity have dedicated AI leaders | Gartner, 2025 |
| High-performing AI companies are twice as likely to report committed senior leaders | McKinsey, 2026 |
Interpretation: Having someone "taking care of AI" is common; having clear responsibility is not. The title matters less than the mandate and responsibilities.
Profile of the Responsible Person
| Requirement | Why |
|---|---|
| Business vision | The agenda must start from business problems |
| Cross-departmental access | Almost every use case crosses departments |
| Executive Team mandate | Without formal authority, every step becomes a negotiation |
| AI fluency | Must evaluate proposals and recognize limits without necessarily being technical |
| Allocated time | An agenda led "in spare time" does not progress |
dooop Framework: The 5 Responsibilities
| Responsibility | What It Does |
|---|---|
| 1. Agenda | Proposes priorities to the board and connects AI to strategy |
| 2. Portfolio | Maintains inventory, monitors initiatives, and recommends continuation, scaling, or stopping |
| 3. Governance | Ensures policy, risk classification, controls, and incident handling, with legal and risk teams |
| 4. Training | Coordinates literacy, manager preparation, and a network of focal points, with HR |
| 5. Reporting | Reports results, risks, and incidents to the board and, if applicable, the board of directors |
dooop Framework: 3 Models by Company Size
| Model | When to Use | How It Works |
|---|---|---|
| A. Combined Responsible | Medium companies, early AI years | A director or senior manager combines the role; a small core team of 2 or 3 people dedicates part-time (IT, legal or risk, HR) |
| B. Dedicated Responsible with Committee | Companies with multiple operational cases | Full-time responsible person; monthly committee with directors from key areas |
| C. AI Office | Large organizations or those with AI central to the business | Dedicated team (product, data, governance, training); common standards; departments with focal points |
These models are stages: many companies start with A and move to B as the portfolio grows. The division of decisions among CEO, IT, and business areas is detailed in How to Align CEO, CIO, CTO, and Business Areas in AI.
Common Mistakes
- Assigning AI to IT by default, without a business owner.
- Appointing without mandate: the responsible person cannot decide anything alone.
- Confusing the agenda owner with the owner of each system: each AI use also needs an owner; see How to Create Clear Responsibilities for AI Systems.
- Creating the role and forgetting executive sponsorship: the responsible person leads but needs the CEO; see How to Build Executive Sponsorship for an AI Agenda.
What to Do Now
- Choose the model appropriate to your size and stage.
- Write the mandate in half a page: responsibilities, authority, allocated time, and reporting line.
- Communicate the appointment to the company, including contact channels.
To structure AI roles and governance, explore the AI Governance area and dooop’s consulting services.
Frequently Asked Questions
Do we need a Chief AI Officer? Not necessarily with that title. What matters is mandate, time, and clear responsibilities. Most AI leaders, according to Gartner, do not have the Chief AI Officer title.
Should the AI responsible person be the same as the data or security officer? It can be, in medium companies, provided they have business vision and time. In larger companies, these functions are usually separate but collaborate.
What about the Data Protection Officer (DPO)? The DPO required by LGPD plays an important role when AI involves personal data, but their function is data protection, not leading the AI agenda.
Also Read
About dooop
dooop is a consultancy based in Caxias do Sul (RS) that prepares leaders and organizations to transform artificial intelligence into business capability through strategy, governance, and training.