AI Initiatives Still Producing Unclear Results

Artificial intelligence has quickly become part of everyday work in the finance function. Teams use it to pull figures from documents, automate payables, draft commentary, and flag anomalies. Yet a nagging question hangs over all that activity: is any of it actually paying off?

According to new Gartner research, the honest answer for most organizations is "we don't know yet." Even as finance AI adoption matures, AI initiatives still producing unclear results remain the norm rather than the exception. For CFOs, that uncertainty changes how AI investment decisions should be made, tracked, and, when necessary, ended.

AI Initiatives Still Producing Unclear Results: What the 2026 Gartner Survey Found

Between January and April 2026, Gartner surveyed 160 senior finance leaders as part of its broader C-Suite AI Survey. The headline number looks encouraging: 55% said their 2025 AI initiatives delivered positive returns overall.

Look closer, though, and the clarity disappears.

Positive Portfolio Returns, Murky AI Project Returns

When leaders assessed the return on AI investment for the most common finance AI use cases one by one, the results were far less tidy.

ROI of AI projects at the individual use-case level

Outcome	All use cases evaluated	Use cases with a clear return (43%)
Unclear return	57%	—
Positive return	18%	42%
Negative return	19%	45%
Break-even	6%	13%

Source: Gartner survey of 160 senior finance function leaders, January–April 2026

In other words, more than half of the AI use cases finance teams are running have no measurable answer on value. And among the minority that can be measured, negative AI project returns slightly outnumber positive ones.

Why AI Portfolio Management Relies on a Few Big Wins

How can most finance leaders report positive returns overall while most individual projects are unclear or losing money? Gartner explains that AI portfolios run on asymmetric returns. Many initiatives fail to recoup their AI project costs, but a handful of successful projects generate enough value to cover those losses and more.

That makes AI portfolio management for CFOs a balancing act. Finance leaders have to keep funding pilots knowing that a meaningful share of that money will become a sunk cost, while making sure the few high-impact bets get the resources they need to scale.

Why AI Initiatives Produce Unclear Results

The survey points to a gap between AI activity and provable AI business value. Several factors drive that gap.

Why AI ROI Is Difficult to Measure

The main challenges in measuring AI ROI tend to show up early, before a project is even launched:

• No baseline. Many teams deploy a tool without first recording how long a task took or what it cost, so there is nothing to compare against.

• Soft benefits. Hours saved on a report rarely show up as a line item unless that time is redeployed to higher-value work.

• Long maturity curves. Complex use cases may not reach full value for more than a year, which means mid-project ROI often reads as negative or unknown.

• Shared infrastructure. Data platforms and integration work support several AI use cases at once, which makes it hard to attribute costs and gains to a single project.

Why AI Projects Fail to Deliver ROI

Beyond measurement, some initiatives simply do not create enough value. Common causes include choosing use cases with weak strategic fit, launching before the data and processes are ready, and lacking the internal skills to adopt the tool fully. Without clear performance targets, these underperforming AI initiatives can run for months before anyone questions them.

How Long Does AI Take to Deliver ROI?

Patience is part of the equation, but how much depends on the use case. Gartner asked finance leaders how long each AI use case takes, or is expected to take, to deliver its full expected value after deployment.

Quick-Win AI Use Cases vs. Long-Horizon Bets

Transactional and process-heavy work, such as data extraction, AP and AR automation, and report creation, typically pays back in about nine to ten months. More analytical work, including forecasting, insight generation, and scenario planning, needs longer development before it delivers. Setting realistic time-to-value expectations for each category keeps leaders from writing off a promising project too early or tolerating a weak one for too long.

Where Finance AI Spending Is Going

Gartner's data also shows how finance AI investment is being allocated. About 35.3% of finance AI spending targets productivity improvements, such as process optimization and fewer employee hours per task. Another 20.9% goes toward cost reduction, including supplier spend and headcount savings.

Smaller slices, each between 5% and 9%, support goals like evolving the enterprise with new products or processes, managing risk and compliance, building operational resilience, growing revenue, transforming the business model, and retaining revenue through better customer service.

Adoption follows the same pattern. The AI use cases finance teams are most likely to be running today are the quick-win ones, while the more strategic ones are mostly still on the roadmap.

AI Productivity vs Business Value

Productivity gains and AI cost savings matter, and they are often the fastest way to show progress. The risk is that they crowd out everything else. Gartner warns CFOs not to let the pull of quick returns sideline the more complex projects that improve decision-making, strengthen risk management, and support revenue growth. Those longer-horizon use cases are often where the outsized wins in an AI portfolio come from.

AI Investment Strategy for CFOs

"AI adoption has reached a point where CFOs must adopt more deliberate portfolio management," said Marco Steecker, Senior Director Analyst in the Gartner Finance practice. The aim, he noted, is not to discourage experimentation but to know where to invest, when to cut underperforming initiatives, and which foundational capabilities to speed up.

Here is how that AI investment discipline translates into practice.

How CFOs Should Evaluate AI Investments

Because so many AI investment returns are uncertain, Gartner recommends that CFOs set AI investment priorities based on three factors rather than projected ROI alone:

• Strategic Alignment: Does the use case support a business priority, not just a task-level efficiency?

• Organizational Readiness: Are the data, processes, and people in place to adopt it?

• Scalability: If it works, can it expand across teams, entities, or processes?

How to Measure AI ROI and Track AI Project Performance

Knowing how to measure return on AI investment starts before deployment. Capture a baseline for time, cost, and error rates. Define the business outcome the project should move, whether that is faster close cycles, better forecast accuracy, or lower supplier spend. Then review AI investment performance against those metrics at set intervals that reflect each use case's expected time to value. Measuring AI project performance on a schedule, and measuring AI business value by outcomes rather than activity, turns vague impressions into evidence the CFO can act on.

When to Stop an AI Project

Disciplined portfolios need an exit plan. Gartner advises attaching performance targets and termination timelines to every initiative. If a project misses its milestones well past its expected payback window, without a clear path to recovery, that is a signal to stop and redirect the budget toward stronger candidates.

Close the AI Literacy Gap

Finding the right talent used to be the top obstacle for AI in finance. With agentic coding tools making AI easier to build and use, low AI literacy has now become the biggest barrier. Gartner suggests giving employees hands-on exposure through project assignments, sandbox experimentation, and short on-the-job activities so teams build the skills and confidence to get more out of every tool.

Getting Value From AI Investments Starts With Discipline

The takeaway from Gartner's research is not that AI in finance is failing. It is that most organizations are investing faster than they are measuring. Portfolios can still come out ahead thanks to a few large wins, but that is not a strategy CFOs can rely on indefinitely.

Getting value from AI investments means treating them like any other capital allocation: set realistic timelines, measure against real baselines, balance quick productivity wins with longer-term business outcomes, invest in AI literacy, and be willing to cut what isn't working. Do that, and AI initiatives can move from unclear results to returns the finance team can prove.