Chief Financial Officers need structured finance AI roadmaps to improve results from artificial intelligence projects, according to guidance presented by Gartner Analyst Ash Mehta at the firm's finance gathering in London.
Data shared by Gartner showed a wide gap between AI activity and reported outcomes in finance teams. In a June 2025 survey of 183 Chief Financial Officers, 84% of finance organisations said they had implemented or were planning to implement AI, yet only 7% reported a high or very high impact.
Mehta argued that the mismatch shows spending alone does not explain success. Instead, finance functions with the strongest results tend to follow a more disciplined process that links AI work to business outcomes.
“Organisations that succeed with AI are not necessarily smarter, luckier or better funded. Rather, they follow a structured and disciplined roadmap that connects finance AI initiatives to business outcomes,” said Mehta, Senior Director Analyst at Gartner.
Gartner's approach has three stages: set a vision and identify current maturity, build a roadmap, and execute and scale use cases.
Set the vision
The first step is to define what an AI-enabled finance function should look like. That vision should answer three questions: the desired end state for finance, how finance will use AI to support enterprise objectives, and what value AI in finance should deliver to the wider business.
This emphasis on vision reflects a broader issue for finance teams that have launched pilots without agreeing on how success will be measured. Without that clarity, organisations can struggle to decide which projects deserve more funding and which should be dropped.
Maturity model
Gartner also urged finance leaders to assess organisational maturity before expanding AI use. Its finance AI maturity model sets out five phases: avoid, experiment, stabilize, expand and transform.
The model covers four workstreams: culture and leadership, strategy and governance, skills and organization, and software and data. Chief Financial Officers can use the framework to identify gaps holding back progress and turn those findings into a sequenced roadmap.
That roadmap should be modular rather than fixed, with priority given to workstreams where finance is lagging. Actions can then be staged across foundation and quick wins, scale and optimization, and innovation and leadership.
Use-case discipline
The final stage focuses on how finance teams select and run AI projects. The roadmap should be translated into a use-case cycle built around identifying, prioritising, executing, scaling and refining projects.
Under that process, finance leaders are advised to create clear intake and approval routes for new ideas. Proposed use cases should include documented objectives, costs and benefits, and should be assessed on net business value, feasibility and scalability.
Mehta said finance teams should avoid pursuing too many projects at once, a common problem in organisations that have rushed to test AI tools across multiple functions.
“If everything is a priority, nothing gets funded. CFOs should identify 3-5 use cases to pilot at a time,” said Mehta.
The recommendation points to a narrower, more selective approach than some finance teams have taken during the recent rush to trial generative AI and automation tools. Limiting the number of pilots can help finance leaders compare outcomes more clearly and decide which use cases to expand.
Gartner also said the roadmap should not be treated as a static document. As business priorities and AI tools change, finance teams should revisit assumptions, update the sequence of actions and scale the projects that show results.
“The strongest roadmaps are living plans, not static documents. Finance leaders should customize the roadmap to their organization, review it as AI capabilities and business priorities evolve, and aggressively scale the use cases that succeed,” said Mehta.