Target operating model: a leader's guide to building one that gets used

Definition: A target operating model (TOM) describes how your business will run once a planned change is done: the processes, people, technology, data, AI and controls that deliver your strategy, and how they fit together.

In short: Most target operating models fail because they're drawn at the top and ignored on the floor. A good one passes the floor test: the people who run the work recognize it, know what changes for them, and helped write it. It gives every layer a named owner, has a place for AI, and ends in a business case you can fund.

I started as a manufacturing engineer, and I still think about operations like a factory floor. If the people on the floor wouldn't recognize your operating model, it isn't a model. It's fiction.

Why does this matter beyond the strategy deck?

Money, risk and control. Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027, citing rising costs, unclear business value and weak risk controls. Each of those is an operating model question in disguise. You can't judge the return on AI without knowing the operating model it plugs into. In regulated operations, the model has to show where a person checks the work and how every action is recorded. And every layer needs a named owner who can say yes or no.

What does a target operating model include?

Eight layers, each answering one question. If you can't fill one in on a single page, you don't yet know how that part of your business will run.

• Services: What do we deliver, and what does good look like to the client?
• Processes: What's the standard way of working, and where are the exceptions?
• People: Who owns each part, and who has authority to decide?
• Technology: Which systems do we keep, drop or connect?
• Data: Which record wins, and who owns it?
• AI: What does AI do alone, what goes to a person, and who owns the rules?
• Controls: How is it recorded, reviewed and stopped if needed?
• Measures: How will we know it's working, and what's it worth?

Don't confuse it with strategy (where you're going), your current operating model (how work runs today) or a roadmap (the order you get there in). You need all four.

What is the floor test?

It's the test I'd put any target operating model through before it goes near a board. Three questions, put to the people who run the work:

1. Do you recognize it? Does it describe how work really happens, exceptions included?
2. Do you know what changes for you on Monday? Can each person say what they'll do differently?
3. Did any of you help write it? Or did it arrive from above?

If the answer to any of them is no, the model will be worked around, however good it looks. Most target operating models fail the floor test, because they're designed at the top, often by outsiders, and delivered as a diagram.

[Add a personal story: a model that failed the floor test, anonymised, and what happened.]

What does a filled-in one look like?

This is TMF Group's, in before and after form. TMF delivers corporate, fund, HR and payroll services across more than 80 jurisdictions.

• Services: From delivery that varied by country, to consistent delivery in every country.
• Processes: From over 900 bespoke processes, none documented, to under 200 in one standard set.
• People: [To add with TMF: how roles and decision rights changed.]
• Technology: From bots working alone with no integrated core, to one orchestration platform and one core technology layer.
• Data: [To add with TMF, if it can be shared.]
• AI: From AI that felt bolted on, and automation that sometimes cost more than it saved, to AI embedded across operations by default.
• Controls: From no visibility of what was being delivered and where, to full visibility.
• Measures: A $43M annual margin improvement.

Every layer changed together. That's the point of the model, and it's why a change in one layer alone rarely sticks.

How does AI change it?

It adds questions the model has to answer: which work AI does, where a person checks it, who owns its rules, and what's recorded. The order matters. TMF retired its scattered automation and standardized its processes before AI became part of how it operates. Utmost Group did the same, streamlining its processes before automating any of them, and cut internal processing time by 80% within six months. Both sorted the model first. AI came second.

How do you build one?

1. Start from where you are. Map current systems, data and processes, including the workarounds, with the people who use them.
2. Design the layers together, with your people. Processes, roles, technology, data, AI and controls change together. The people who'll live with it should help write it, so they carry it through.
3. Put AI in the model. Decide where it works and where a person checks.
4. Build the business case. Show what the change is worth and what standing still costs.
5. Sequence it, and give every layer an owner. Separate quick wins from longer-term work, and review it, or it drifts back to the workarounds.

Sign-off checklist

1. Has it passed the floor test?
2. Who owns each layer, by name?
3. Where does AI work, and where does a person check it?
4. What will it cost, what is it worth, and what does standing still cost?
5. How will we know in six months whether it's working?

If your team can't answer these, it isn't ready to fund. A target operating model is one of the four things our Health Check delivers, and it starts with a free discovery call.

Key takeaways

• A target operating model is how your business will run once change is done, across eight layers including AI and controls.
• Put it through the floor test: recognize it, know what changes on Monday, helped write it.
• Sort the model first, and AI second, and end it in a business case you can fund.

FAQs

What is a target operating model?

A target operating model is a description of how a business will operate once a planned change is done, covering its processes, people, technology, data, AI and controls, and how they fit together to deliver its strategy.

What is the difference between an operating model and a target operating model?

An operating model describes how a business runs today. A target operating model describes how it will run after the planned change.

What is the difference between a target operating model and a strategy?

A strategy sets where the business is going and why. A target operating model describes how the business needs to run to get there.

What is an AI operating model?

It's the part of a target operating model that describes how AI will work in the business: what it does, where a person checks it, who owns its rules and what's recorded.

What is the floor test?

It's a three-question check for a target operating model, put to the people who run the work: do you recognize it, do you know what changes for you on Monday, and did you help write it.

How long does it take to build one?

It depends on the scope. Ours comes out of a Health Check that takes three hours of your time and three weeks of ours.

Sources: Gartner, June 2025, on the cancellation of agentic AI projects. [Add link.]