From AI Enthusiasm to Enterprise Value

In our last article, we concluded with a simple but important observation: AI is only as effective as the context that surrounds it. The most sophisticated model, the most advanced agent, and the most compelling demonstration can all produce poor outcomes if they operate without a foundation of trusted data, governance, and business understanding.

When I speak with stakeholders ranging from CFOs and CIOs to first-line finance managers, I often find that the conversation about AI quickly shifts away from technology itself. Very few executives ask me about large language models, vector databases, or agent frameworks. Instead, they ask a much simpler question: "Why should this matter to me?" The answer not only needs to be convincing to the person asking the question, but also, at large, answer their stakeholder's query about: "What are we doing with AI and how is it adding value?" This question rings loud in the board rooms and investor summits.

As I write this article, the AI wave is gaining tsunami-like ferocity. Last week I found myself in San Francisco for the first time in over a year. It was impossible to miss the AI renaissance unfolding around the city. Billboards, advertisements, conferences, and conversations all pointed toward the same message: AI is rapidly moving from experimentation to enterprise execution.

AI is not simply a productivity tool. Its real potential lies in challenging the assumptions behind how organizations plan, report, reconcile, allocate, and make decisions. On this trip to SFO, I was reading "The AI Driven Leader" from Geoff Woods. There is an entire section dedicated to "The High Price of the Wrong Questions" which talks about how asking wrong questions (think wrong prompts) of AI can magnify our mistimed and misplaced biases and assumptions. The section elementally argues that we need to use the undisputed analytical capabilities of AI with human-led intuition in order to make impactful strategic decisions for the near and long term. So I ask myself and the readers: "What are some areas where the leaders are asking for AI's help?"

To get us started, here is some of what I have received from finance leaders in my deliberations:

• I want to detect anomalies in my business data faster.
• When I detect anomalies, I want us to explain the variance automatically.
• I want to forecast better.
• I want to reconcile smarter and faster.
• I want to allocate more transparently and have the ability to conduct agile what-ifs.
• I want to close more efficiently with automation for journaling where possible.
• I want to manage my enterprise charts efficiently and be aligned with industry standards and forward-looking reporting capabilities.
• I want to enable my organization with fast and accurate information.
• I want to self-service reporting insights.

While these requests sound like AI problems on the surface, they are often readiness problems underneath. Before discussing technology, it is important to understand the foundational capabilities that determine whether AI produces sustainable business value or simply accelerates existing inefficiencies. Across my conversations with finance leaders, the following themes consistently emerge.

The foundations that decide the outcome

The encouraging news is that many of these requests are no longer aspirational. Oracle has been steadily embedding AI capabilities directly into the EPM platform for years. Predictive Planning, Intelligent Performance Management (IPM), machine learning-based transaction matching, AI-driven narrative generation, and emerging conversational assistants all represent targeted responses to the challenges finance leaders are articulating.

The more important observation, however, is that these capabilities are not standalone AI solutions. Each one relies on varying degrees of data maturity, process discipline, governance, and organizational trust. Without those foundational elements, the technology may function perfectly while the business outcome still falls short.

To illustrate this relationship, I mapped common finance leadership objectives to Oracle EPM embedded AI capabilities and the contextual requirements necessary to generate meaningful business value.

A contextual map of Oracle EPM embedded AI

Start with the business problem

At this point, we have introduced the three components that matter most:

• Who is asking for AI?
• What outcomes are they seeking?
• What business, data, and governance requirements must exist for AI to create value?

Only after answering those questions should we begin discussing technology.

Too often organizations start with features and work backward toward a business problem. The more successful implementations I have seen begin with the business problem, establish the necessary foundations, and then selectively deploy AI where it can create measurable impact.

The matrix above is not intended to be a product catalog. It is intended to serve as a contextual map that connects Oracle EPM's embedded AI capabilities to real finance outcomes and the organizational disciplines required to support them.

In the next articles, I will take several of these use cases and unpack them individually — examining both the business process architecture and technical architecture required to move from AI enthusiasm to enterprise value.