Business Owners for AI Projects: Why IT-Only Pilots Stall

The IT-led pilot usually goes like this. IT spots a use case it’s sure will help the sales team, builds a clever prototype, and demos it. The sales team is polite. A few people try it. Two weeks later, usage has quietly dropped to zero. When IT asks why, the honest answer is that sales never asked for it, never said what success would look like, and never changed how they work to make room for it.

Nothing was wrong with the technology. What was missing was an AI project business owner: someone on the business side who wanted the result, defined it, and was accountable for getting it. In my experience, that one missing role explains more stalled AI pilots than every technical problem combined.

Why IT-only pilots stall

IT is good at building and running systems. On an AI project, IT is poorly placed to do five other things that decide whether it succeeds:

• Define value. IT can guess what would help the business. Only the business knows which outcome actually matters and how much.
• Change the process. AI rarely delivers value by being bolted onto an unchanged process. Someone has to redesign how the work gets done, and IT doesn’t own other departments’ processes.
• Drive adoption. People adopt tools their own manager expects them to use. An IT request carries less weight than a department head’s.
• Fund production. Pilots can run on IT’s budget. Production usually needs business budget, and business leaders fund what they own.
• Fight for it. When priorities shift, a project with no business champion is the first thing cut.

Two roles, at two levels

Business ownership is needed at two levels, and teams often have one without the other.

The executive sponsor works at the portfolio level. They tie AI to the organization’s goals, secure funding, remove obstacles between departments, and make the calls that are above any single project. Usually a member of the leadership team, and ideally not the head of IT, because IT sponsoring its own initiatives doesn’t carry the same weight.

The business owner works at the project level. They define the success metric, sign off on the baseline, provide pilot users and their time, own the process changes, and decide whether the pilot goes to production. Usually the head of the team whose work the AI changes.

IT’s role alongside them is essential and different: judging feasibility, building, securing, supporting, and reporting the numbers honestly.

Here’s how the assessment asks the question about the first of those roles, and the 0 to 4 ladder I score it against:

U3. What executive sponsorship and funding does AI have?

0. No executive interest
1. Interest, but no sponsor or budget
2. Named executive sponsor, no dedicated budget
3. Sponsor plus funded pilots
4. AI is in the strategic plan with multi-year funding

Level 1 is extremely common: leadership is interested in AI, talks about it at meetings, and hasn’t named anyone accountable or set aside money. Level 2 is progress, but a sponsor without budget can encourage and not much else. Level 3, a sponsor plus funded pilots, is where projects start reaching production.

The asset: a one-page business owner agreement

Before any pilot starts, fill this in and get it signed. It takes thirty minutes and prevents most of the ways pilots drift.

The use case: one sentence, from your use-case register.

Business owner: name and role.

Executive sponsor: name and role.

The business owner commits to:

• Defining the success metric and the target that would justify production.
• Signing off on the baseline measurements before the pilot starts.
• Providing a named group of pilot users, with a set amount of time each week for the pilot.
• A 15-minute weekly check-in with IT during the pilot.
• Owning any changes to the team’s process that the pilot requires.
• Making the go, extend, or stop decision by a fixed date.

IT commits to:

• Building and securing the solution to the organization’s standards.
• Supporting pilot users during the pilot.
• Reporting the agreed metrics weekly, including the ones that look bad.

Decision date: the day the business owner will decide.

The metrics side of the agreement is covered in detail in baseline before you build.

A decision rule

No business owner, no pilot. If nobody on the business side will sign the agreement, the use case stays in the register as an idea. That feels slow. It’s much faster than building something nobody will use.

How to recruit a business owner

• Start with whoever complains most. The person most frustrated by a slow process is the one most likely to invest in fixing it.
• Show them their own numbers. “Your team spends about this many hours a month on this” is more persuasive than any AI demo.
• Offer to take something off their plate. Frame the pilot as removing work, not adding it.
• Keep the first commitment small. A few pilot users for six weeks is an easy yes. A department-wide rollout isn’t.
• Give them the credit. When the pilot succeeds, the business owner presents the results. That’s how you recruit the next one.

How to get an executive sponsor

Don’t ask an executive to “champion AI.” It’s vague, and busy leaders decline vague requests. Instead, connect AI to a goal they’ve already stated, such as reducing cost to serve, speeding up a process customers complain about, or handling growth without adding headcount. Bring two or three use cases from the register, each with a business owner already lined up. Then ask for specific things: a modest budget for pilots, thirty minutes a month, and the authority to make cross-department decisions. If you’re the IT leader making this case, answering “what’s our AI strategy?” covers how to frame the larger conversation.

What moving up one level looks like

From 0 or 1 to 2: name an executive sponsor, even without a budget, and give them the use-case register to review. From 2 to 3: ask the sponsor for a defined pilot budget, and run the first two pilots under signed business owner agreements. From 3 to 4: take the measured results from those pilots into the annual planning cycle, so AI appears in the strategic plan with funding that spans more than one year. Each step builds on evidence from the one before, which is why the measurement discipline matters so much. A sponsor who has seen two pilots deliver measured results argues for multi-year funding far more convincingly than one who has only seen demos.

Mistakes I see at this stage

The CIO as sponsor for everything. It’s natural, and it sends the message that AI is an IT project. Find a sponsor on the business side.

Owners in name only. A name on a slide who never attends a check-in isn’t an owner. The weekly check-in in the agreement is how you find out early.

A sponsor with no budget authority. Encouragement doesn’t fund production. Make sure the sponsor can actually allocate money.

IT redesigning other teams’ processes. IT can suggest how work might change. The business owner has to decide and lead it.

Where does your team actually stand?

Executive sponsorship and funding is one of 24 questions in the AI Readiness assessment, which covers six dimensions: data, security, infrastructure, skills, use cases, and governance. The free version is 10 questions and gives you a score in a few minutes.

Get your free AI Readiness Score →

Want to see what the full assessment covers first? Flip through a complete 38-page sample report.

Related guides

• Turn Your Painful-Process List Into an AI Use-Case Register
• Baseline Before You Build: Metrics That Make AI Pilots Fundable
• The IT Director’s Guide to Answering “What’s Our AI Strategy?”
• The 6-Dimension AI Readiness Framework, Explained
• The AI Readiness Checklist: 24 Questions to Answer Before Spending a Dollar