AI in finance has moved from discovery to application almost as quickly as the underlying models themselves have evolved. Finance professionals are already putting AI to work to improve communication, automate reporting, build financial models and support decisions in real time.

AI adoption in finance is unfolding along two dimensions:

1. Scope: This ranges from individual assistants that help people work faster to enterprise AI embedded in business processes and decision-making.
2. Sophistication: This ranges from automation that speeds up routine tasks to analytical reasoning that helps evaluate scenarios, generate insights and support decisions.

Together, these two dimensions create four broad categories of AI use in finance:

1. Individual Productivity
2. Process Automation
3. Augmented Finance Work
4. Decision Intelligence & Agentic Finance

Each of these categories represents a different combination of scope and sophistication.

Viewed through the lens of these four categories of AI use, below is a look at how the finance leaders of AFP’s Emeritus FP&A Advisory Council are using AI in the real world.

AI in Finance Certificate

The No Code AI for Finance Certificate program is built for finance professionals who want real, usable AI skills today. Learn to implement AI responsibly, transparently and in a way that strengthens decision-making across your organization.

Learn More

1) AI as a Better Assistant

The most accessible use of AI is as an everyday assistant for individual users. Individuals at this stage are using AI models to improve the clarity, tone and speed of their written and verbal communication. What they’ve discovered is that the applications allow for faster output with no structural change to finance workflows.

Common use cases include:

• Drafting and refining emails and presentations
• Improving presentations and board commentary
• Adjusting tone and tailoring materials to CFOs and CEOs
• Anticipating executive reactions based on previous feedback
• Conducting basic research and summarizing information
• Preparing for meetings and interviews
• Tracking tasks, commitments and issues requiring their attention
• Simulating interview questions and getting feedback on answers

The end goal with all of these use cases is, as one finance leader put it, to “get it right the first time rather than having to go through multiple revision cycles.”

2) AI as an Automation Layer

When AI is used to improve repeatable processes across multiple people in the enterprise, the tool becomes part of the workflow, supporting data preparation, reporting, analytics and distribution.

Common use cases include:

• Cleaning and combining data from multiple sources
• Automating and refreshing report generation
• Building and distributing dashboards
• Analyzing large financial datasets
• Identifying anomalies, duplication and spending trends
• Automating recurring communications

Using AI capabilities embedded in ERP, EPM and treasury systems

One finance leader said they used AI to examine more than 250,000 general-ledger data points across client accounts. The AI tool can surface trends, risks, duplicate activity and subscription spending that would be difficult for a person to identify by a line-by-line review.

It’s still up to the finance professional, however, to determine whether the findings are meaningful and how they should be communicated to the client. AI can expand the volume of information for examination, but it can’t interpret it.

Reporting is another natural entry point for AI adoption. One FP&A team has used AI-assisted tools to automate the production and distribution of P&L reports, which has improved consistency and reduced the number of routine questions from other business units.

For a smaller finance organization without a dedicated analytics or technology team, vibe coding offers a practical way to prototype and design dashboards. Users can describe the output they want (columns, colors, visuals) in ordinary language, and AI creates the underlying tool or application, saving them a significant amount of time.

3) AI as a Thought Partner and Analyst

For analytical work, finance leaders use AI to help build models, test assumptions, identify weaknesses and accelerate forecasting and planning.

Common use cases include:

• Building forecast and valuation models
• Developing scenario analyses and simulations
• Critiquing assumptions and model logic
• Identifying risks, gaps and blind spots
• Performing multivariate and predictive analysis
• Drafting analytical narratives for senior leaders

Supporting training and development for junior analysts

For one CFO at a smaller organization, AI is filling part of the capacity gap created by not having the budget for a full analyst team. The technology helped build the company’s first forecast model and supported the development of pitch decks and meeting materials. The CFO still reviews the work, corrects formula errors and refines the assumptions.

Another finance professional used AI to build a customer lifetime value model, and then asked it to critique the model, challenge the assumptions and recommend improvements. What may have previously required several weeks to do was completed in approximately 15 hours.

The example above highlights one of the most valuable roles AI can play in FP&A: not simply producing an answer but challenging the thinking behind it. Users can ask the AI tool to identify gaps in logic, test drivers, suggest alternative assumptions or point out risks that may have been overlooked.

4) AI as a Strategic Copilot

When AI is placed inside the decision-making process, the tool helps simulate business processes, pressure-test strategy and surface risks during the decision-making process.

Common use cases include:

• Running real-time simulations during meetings
• Pressure-testing strategic recommendations
• Identifying risks and gaps in business logic
• Supporting strategic planning and board preparation
• Capturing meeting insights and recommending actions
• Exploring future finance operating models

Helping leaders evaluate organizational design and priorities

In a meeting with a CFO, one finance consultant used AI to build a live simulation of three-way matching within a procure-to-pay process. Ultimately, the tool helped identify KPIs and value drivers that AI agents could surface. What might once have taken two or three days and required several employees was produced during the meeting.

Other finance leaders are leveraging AI to challenge strategic recommendations before presenting them to senior management. One example is using AI to evaluate an organization’s long-term physical footprint, including how many locations to keep operational, which markets it should enter and where it might need to reduce its presence. Critically, AI is not making the final decision; it is helping refine the recommendation.

These use cases point toward a future in which AI does more than respond to isolated prompts. It instead regularly participates in workflows, monitors information, identifies issues and recommends actions. Finance can then shift from producing every analysis manually to orchestrating the decision process.

Human Judgment Is Still the Critical Control Point

AI has its limitations. Significant ones. Finance professionals reported that AI use has resulted in formula errors, false positives and models that didn’t work as intended, proving that AI-generated work requires significant quality assurance.

Data security is another serious consideration. Confidential information, personally identifiable information and other sensitive data should never be placed into uncontrolled tools. And it’s important to address the risk of “shadow IT” — when employees build AI-supported processes the technology function doesn’t understand or oversee.

The risk can be mitigated to a degree through documentation. Teams should record the instructions, data sources, validation procedures and workflow behind AI-generated tools so the work can be reviewed, transferred and repeated.

It’s also important to select your use cases carefully. A low-risk dashboard prototype or analysis of a sanitized ledger extract is a good place to begin experimenting. In contrast, a process involving sensitive data, regulatory requirements or material financial decisions requires a more controlled environment.

Those using AI most effectively aren’t treating it as infallible or allowing it to make independent decisions. Even at the level of an individual assistant, the quality of AI’s output depends largely on the financial intelligence built into the prompt — the financial knowledge, business understanding and judgment that finance professionals use to connect numbers to strategy, understand what the numbers mean and determine how they should shape what an organization does next.

AI can generate analysis, forecasts and recommendations, but it doesn't inherently know the business context, key value drivers, strategic priorities, risk appetite or decision criteria — the user has to supply that judgment.

The most effective finance professionals aren't simply asking AI questions; they're embedding business objectives, assumptions, constraints and performance drivers directly into their prompts. That's what allows AI to augment financial intelligence rather than attempt to replace it — expanding the range of questions finance professionals are able to explore and strengthening the analysis behind their own expert judgment.

Learn More About AI in Finance

Learn how to position your AI project, make tactical decisions about AI use in finance and lay the foundation for a successful AI program. Fill out the form below to download the complimentary AFP FP&A Guide to AI-Powered Finance.