Key takeaways
-
Oracle Intelligent Performance Management (IPM) brings embedded AI into Oracle Fusion Cloud EPM, helping finance teams improve forecasts and insights without the need for specialized technical help.
-
A simple four-phase approach—use case selection, pilot, refinement, and rollout—helps organizations start small with IPM, achieve early wins, and expand confidently.
-
Real-world applications include OPEX and revenue planning, where IPM improves prediction accuracy, surfaces key trends, and reduces the need for manual analysis.
We’ve heard some customers say that they’re eager to get started with AI in Oracle Fusion Cloud Enterprise Performance Management (EPM) but don’t know how to begin. Implementing AI features doesn’t have to be a massive undertaking. Often the most effective way is to start small and scale up. If this sounds attractive, Oracle Intelligent Performance Management (IPM) presents a great opportunity.
IPM is a set of AI-enabled features embedded in Oracle Cloud EPM. It uses AI to automate data analysis, provide insights, and improve decision making for finance and operations teams. Among other things, IPM can help improve the accuracy of financial forecasts without the need for data scientists. To get started, you can follow a simple, four-phase process: select a use case, run a pilot, refine the model, and roll out the capability.
Phase 1: Use case selection
Begin by identifying a use case that will help you understand the IPM setup while you score an early win. When selecting a use case, consider the following:
- Impact: Choose a use case that will help you understand IPM predictions, gain confidence, and quickly deliver something valuable (consider ones that would positively impact KPIs your company cares about already). Eventually you can expand to use cases that have a more material impact on the business
- Repetitive: Begin with a use case that includes recurring activities
- Ease of configuration: Start with a simple use case that’s straightforward to configure.
- Data availability: Ensure that the necessary data is already available in Oracle Cloud EPM. For meaningful predictions and insights, you should have at least two years of historical data.
For example, using IPM to generate operational expenditure (OPEX) predictions is a great way to get started.
Phase 2: Pilot
Once you’ve selected a use case, set up a small-scale pilot. The pilot team should include key stakeholders like a finance user, an EPM admin, and a project sponsor. It’s important to note that you probably don’t need a data scientist or systems integrator to turn on AI in EPM. More likely, your team can run a successful proof-of-concept project within a business unit or entity to test AI in an existing process without advanced technical help. The goal is to define what success looks like, such as improved accuracy, speed, or user satisfaction.
Phase 3: Refinement
The next step is to refine how you use IPM by exploring advanced options to improve accuracy. This involves considering how the new AI-driven process affects existing workflows, as well as adjusting IPM settings. For example, you can use IPM’s configuration wizard to access advanced tuning options that let you refine predictions. For example, by making use of multivariate predictions—which use multiple drivers to train and derive predictions—you can change how IPM treats outliers and missing values. Or you can use the configuration wizard to adjust the model’s sensitivity to historical events like promotions.
Even though the primary goal of this phase is to score early wins and refine how you use IPM, you’ll gain valuable experience, too. As you become familiar with advanced options, you’ll gain the expertise and insights needed to help you expand the use of IPM to additional, more impactful use cases.
Phase 4: End-user rollout and steady-state care
The final phase is about change management, user education, and the ongoing evolution of business processes. This involves training end users to interpret and use IPM’s predictions and insights. It’s also critical to monitor the model’s performance and continue to refine business processes as the organization’s AI capabilities mature.
Two examples
Here are two examples showing how AI-powered features in Oracle IPM can help.
- OPEX planning: In a traditional EPM environment, forecasting OPEX may be difficult when historical data for specific sub-expenses, such as travel or utilities, is incomplete. With IPM, you can generate predictions at any level in your expense hierarchy. This helps improve topline OPEX forecasting without requiring granular data. Also, IPM can then surface key expense trends at all levels, reducing the need for manual analysis.
- Sales and revenue planning: You can use IPM to explore advanced sales and revenue predictions. For example, sales predictions may have correlations to historical sales volumes, promotion expenses, industry volumes, as well as external economic indicators such as gross domestic product. Using IPM’s advanced predictions, you can model these drivers for higher accuracy.
Conclusion and next steps
Integrating AI into your EPM processes can deliver quick wins and help you gain valuable experience—and it may not be as hard as you think. By starting with a clear use case and following a phased approach, you can build a solid foundation for future success.
To learn more, check out IPM Best Practices for AI Adoption in EPM on Oracle Cloud Customer Connect.
Related posts you might like
If you’re an Oracle customer and want to get new stories from The Fusion Insider by email, sign up for Oracle Cloud Customer Connect. If you’re an Oracle Partner and want to learn more, visit the Oracle Partner Community.