Spending on enterprise Gen AI tools continues to accelerate. And yet, most executive teams still cannot point to clear, measurable ROI. One core obstacle is that many organizations manage AI investments primarily as technology investments rather than as business transformations.
Much like with a major ERP implementation, a company’s earliest AI initiatives commonly originate within IT, with investments tracked along other IT line items such as software applications, cloud infrastructure, and hardware upgrades. The problem with managing AI as IT line item is that it creates an accountability void. IT budgets exist to buy technology solutions, keep systems running, and support users, not to transform how business teams work day-to-day.
To unlock the real financial upside of AI, AI initiatives must be owned by BU leaders held accountable for ROI, not treated as just another technology purchase.
Why the IT budget model is flawed
CIOs and IT organizations are well-equipped to procure, integrate, and secure technology. They are not, however, positioned to drive change in how operational or functional teams execute their daily work.
When AI is evaluated primarily as an IT purchase, success is defined primarily in terms of technical milestones: live system launch, license activation, security requirements established. Business success, by contrast, is measured by concrete results, such as higher margins, lower expenses, or greater throughput.
Without a functional business leader held accountable for AI adoption and returns, the outcome is often underutilization: new tools have been fully installed and made available, but day-to-day behavior has changed very little.
Aligning executive accountability
Establishing financial return requires defining operational guardrails across three primary executive roles:
1.) The Chief Financial Officer
The CFO controls capital allocation and insists that the full expense profile—licenses, training, integration, change management, and ongoing support—is captured. They hold actual financial results against the original business case before releasing any additional funds.
2.) The Business Unit Leader
Any functional leader (COO, CMO, VP of Supply Chain) must maintain ultimate accountability for the AI business case within their domain. While IT delivers technology, this leader owns the operational adoption and the conversion of productivity gains into actual financial outcomes: reduced expense, additional revenue, or greater capacity that is deliberately redeployed.
Note: Productivity gains (including efficiency improvements and time savings) do not automatically equal financial returns. Saving employees time only creates business value when time saved is deliberately redeployed toward lower spending or greater revenue generation.
3.) The Chief Information Officer
The CIO owns the tech stack, systems integration, data security, and platform performance. They partner with BU leaders to assess technical feasibility and ensure the AI capability can be embedded into existing workflows.
Reframing AI budgeting: a practical approach
Defining the domain of these three roles is important, but real leverage comes from considering them in a shared decision-making and capital allocation structure. Instead of approving large spending upfront, then hoping for results, organizations should manage AI spending through a staged, outcome-based model.
Stage 1: Upfront alignment
Before any meaningful spending, the business unit leader, CIO, and CFO must align on a clear plan. This includes agreeing on the exact business problem they are trying to solve, naming who owns the outcome, establishing the baseline against which progress will be measured, and mapping the full cost (beyond just the AI tool’s sticker price). With that foundation in place, start with tightly scoped pilots that have explicit criteria for success rather than committing large spending upfront.
Stage 2: Proof before additional funding
Once the pilot is up and running, additional funding may be unlocked by meeting performance milestones. Before funding wider rollout, leadership must see evidence that employees have actually adopted the tool and that output quality meets established standards. They must also make sure the business leader can demonstrate how any productivity gains are converting into lower expenses or higher revenue. Clear, measurable proof in a real operating environment becomes the gate for the next round of funding.
It’s really not that complex
At its core, this approach is simply good business discipline. For nearly any other significant business initiative, organizations would assign a business owner, establish a clear baseline, start small, and require evidence of return before allocating more funding. AI shouldn’t get a pass on these standard business practices simply because the technology is newer.
