AI operating model: who decides, who builds and who is accountable
An operating model answers three questions: who chooses what gets built, who builds and runs it, and who answers for the result. Without it, every department buys its own tool and nobody owns the risk.
An AI operating model sets how a company decides on, builds, runs and governs AI. For most small and mid-size companies the right shape is hub-and-spoke: a small central group sets standards, approves tools and runs shared platforms, while each business unit owns its use cases and results. A steering committee of executives sets priorities and settles disputes.
Microsoft 365
Google WorkspaceSalesforce
ServiceNow
Slack
NIST
How to read this chart
The scores are our planning judgement for a typical small or mid-size company: they show where effort or risk tends to concentrate, not measured data. Replace them with your own once you have looked at your company.
Centralised, hub-and-spoke or federated?
Centralised is fastest to control and slowest to scale. Federated scales fastest and duplicates the most. Hub-and-spoke gives a small company standards without a bottleneck.
- Centralised: one team builds everything. Fits the first six to twelve months or a company with one main business line.
- Hub-and-spoke: a small hub sets rules and runs shared pieces; units own use cases. Fits most companies once two or three teams are active.
- Federated: units run their own programs under light common rules. Fits groups of distinct businesses, such as a holding company.
What does the steering committee do?
It sets priorities, approves funding, removes blockers and reviews results. It does not choose tools or review prompts.
- Members: the executive sponsor, finance, operations, technology, legal or risk, and the heads of the units with live use cases.
- It meets monthly while the program is young and quarterly once it is steady.
- Every meeting reviews the same short scorecard: results against target, spend against budget, incidents and the next decisions.
Do you need an AI centre of excellence?
Only once the work justifies it. A centre of excellence is the hub with a name: the people who keep standards, shared tools and training in one place. In a small company it can be two people part-time.
- Start with a named AI lead and one technical person.
- Add a training owner when more than a few teams are using the tools.
- Keep delivery close to the business; a centre that builds everything becomes a queue.
The NIST AI Risk Management Framework's Govern function asks for the same things in plainer terms: clear roles, accountability and policies that are actually followed.
Questions executives ask about this
Who should chair the AI steering committee?
The executive who owns the business results AI is meant to move, usually the CEO or COO. The technology lead should be a member, not the chair, so priorities stay tied to the business.
How big should the central group be?
As small as it can be while still answering requests quickly. Many mid-size companies start with one lead and one builder, and add capacity when the request queue grows.
Where does IT fit?
IT owns identity, security, integrations and the platforms. The business owns the use cases. The operating model exists so those two do not wait on each other.
Can an outside firm act as the hub?
Yes, for a time. A fractional AI lead or a consultancy can set standards and run shared pieces while the company builds its own capability, as long as an internal executive stays accountable.
Primary sources to check against
AI Risk Management Framework, AI RMF 1.0 (National Institute of Standards and Technology)
ISO/IEC 42001:2023, AI management systems (International Organization for Standardization)
Frameworks, laws and product details change. Confirm the current text with each publisher and with your own counsel. Consultants in a Box is an independent consultancy; this page is not published or endorsed by any organisation named. Names and logos are trademarks of their owners.
Related guides
Chief AI officer
Someone has to own AI across the company. That does not have to be a new full-time executive. The decision is about accountability first and the title second.
Read the guideAI governance frameworks
You do not need to adopt every framework. You need to know which ones are voluntary, which one is law where you operate, and how to build one set of controls that satisfies them.
Read the guideBoard oversight of AI
A board does not need to understand how a model works. It needs to know where the company uses AI, what could go wrong, who is accountable and whether the investment is paying back.
Read the guideAll AI for executives guides
The full leadership playbook, from readiness to agents.
See the hubWant to talk this through for your company?
Book a call and we will go through where you are, what to decide first and who should own it. Or start with the free business analysis.
