Who runs AI in your organization?

Start with clear roles and responsibilities

If you’ve already set your AI strategy, the next step is turning it into a program your organization can put into action. That means deciding how the work gets done and who owns what. IT may own the technology, while business leaders own the outcomes. Security and compliance set guardrails, finance influences funding, and AI leaders and Champions often connect these groups without having final authority.

Getting those roles right matters. Microsoft Work Trend Index research found that organizational factors, including culture, manager support, and talent practices, account for roughly two-thirds of AI’s reported impact. In other words, realizing the value of AI depends on more than the technology itself. It also depends on how your organization puts it to work.

A practical operating model for your AI program starts by answering three questions: Who makes the decisions? Who runs the program? Who funds it?

Decide who runs your organization

AI decisions rarely belong to one person or team. Start by mapping the decisions that need to be made for your AI program. These might include:

For each decision, identify who’s responsible for the work, who’s ultimately accountable, who should be consulted, and who needs to be informed. The goal isn’t to give one person authority over everything, but to establish clear roles before a decision is needed.

Clarify who funds the work

Funding an AI program is another area where AI leaders often influence decisions without owning them.

Your AI strategy provides a starting point for the funding conversation. It defines the desired business outcome, prioritized workflows, current baselines, and 90-day targets you can use to build the business case with your finance team.

Bring finance partners in early so they can provide input on how costs and expected returns are evaluated, rather than waiting until the end to ask for approval.

Your finance and FinOps partners can also guide decisions about cost management, license optimization, forecasting, and how AI spend is allocated across business units.

For guidance on forecasting and planning costs, read Forecast and plan Microsoft Copilot spend across services.

Assign ownership for AI agents

The number of active agents is growing rapidly across the Microsoft 365 ecosystem, according to Work Trend Index research. As organizations expand their use of agents, they also need to establish who owns them and is accountable for their activity.

For each agent:

• Give it a named owner.
• Define its permissions.
• Make its activity reviewable.
• Establish who is accountable for its outputs.

Agent ownership doesn’t need to be treated as a separate organizational matter. Include agents in the same conversations about ownership, governance, and decision-making that apply to the rest of your AI program.

Put your operating model in writing

Once you’ve clarified who decides, who runs the program, and who funds it, put those decisions into writing so everyone knows who owns what.

A RACI matrix is a project management tool that gives you a practical way to clarify the roles people play in key decisions: who’s responsible, accountable, consulted, and informed.

Use this RACI template to map these roles to those decisions in your AI program. Complete it with your executive sponsor and IT counterpart, then revisit it as your program evolves.