Note: this is the fourth post in a 5-part blog series about the relationship between Oracle and AI, specifically in the EPM space.
Welcome back! Just a quick warning, this is going to be a biased post. Oracle EPM reporting is my sweet spot and where I focus my consulting work. I’ve known the Oracle EPM reporting team for many years and I speak to them regularly. It’s a fun group! 😄
I reached out directly to the Oracle EPM product managers currently in charge of the broad spectrum of solutions we call “EPM reporting.” Al Marciante, Tom Lefebvre, and Dave Roberts work specifically with Narrative Reporting, amongst other EPM solution sets. You’ve most likely heard their voices over a Cloud Customer Connect webinar or sat in the audience as they presented at an EPM conference. They always do a very thorough job of explaining new features and answering questions about the Oracle EPM reporting stack.
EPM reporting is one of the areas, much like Analytics, where I envision several exciting possibilities when it comes to Ai. And Oracle’s vision, to be revealed later in this post, is on par with where I would hope Oracle EPM reporting would go.
But first, a terminology lesson…
Since EPM reporting falls under the Cloud EPM stack, this means Ai-related technology is referred to as “IPM” or intelligent performance management.
In addition, you will see the term “IPA”, or intelligent process automation. To be clear, this is not a type of beer. 😂
IPA combines robotic process automation (RPA) with machine learning (ML) to manage and automate EPM processes. It uses software bots to perform rule-based operations, learn from situations, make decisions, perform visual interpretations, and adapt. In the Cloud EPM world, this correlates to functionality like reconciling balance sheets and automating consolidations.
What Ai technology exists in Oracle EPM reporting today?
Let’s start by discussing what, if any, Ai exists in the Oracle EPM reporting tools today.
Although there’s a bigger and more cohesive vision for EPM Reporting + IPM in the future, the team mentioned that within the Oracle EPM reporting stack, not much of traditional Ai exists currently. Instead, they leverage the data outputs of IPM.
Wait a second. But what about…? Yes, there have been Monte Carlo simulations in Finance reporting. There is also the concept of “conditional text” in the Reports tool, also called “narrative generation.” However, as it was explained to me, conditional text is not classified as Ai or IPM. It falls under the classification of IPA instead, as it helps automate the addition of textual notes via conditional formatting rules.
And if you’re curious about some of the cool features discussed for Task Manager, another area Tom works on, this technically falls under IPA as well.
So, everyone, just be patient and keep reading. The details of what they’re working on and when it will be released is revealed later in this post. 😉
How exactly does the process work when creating new Ai features and innovations in Oracle EPM reporting?
There are so many teams at Oracle and with overlap in the Ai space. So how exactly do the teams work together when new Ai innovations are requested by the Oracle EPM team? This is something I’ve been curious about but haven’t dug into yet.
As Al explained, “the name ‘Oracle’ implied years ago they would be in the Ai space.” So, it’s not surprising that Oracle finds itself already becoming quickly established as leaders in this technology arena.
The EPM team works with the FusionAi team, which is a development center at Oracle. They are a centralized team that facilitates and coordinates the utilization of GenAi in SaaS applications. Teams at Oracle register use cases with them internally to develop new features in their SaaS solutions. The FusionAi team is responsible for working with data centers and other groups to ensure a smooth process.
As Al explained, the Oracle EPM team has a lot of use cases, more than many other teams, which can generate stress to the Ai service and the EPM team. The FusionAi team alleviates some of the havoc by helping Al’s team form the right questions to get the expected results, as well as provide quality assurance. In addition, without this team, internal costing would work differently—the EPM team would have to maintain GPUs and deal with utilities consumption and maintenance costs.
As you can tell, FusionAi provides a lot of services. There’s a lot of R&D they’re working on, and they provide the critical focus and attention on security when things go live. Approvals and success criteria are required before customers ever see these new innovations and, on top of everything else, the FusionAi team helps manage those.
As this topic progressed, Al also made it clear that Oracle doesn’t train models with customer data, but they do use large language models so Ai can understand their Finance and EPM use cases. These use cases require learning arithmetic and math. Due to this, they have a relationship with Cohere. Cohere is a 3rd party company who specializes in large language models to provide natural language processing (“NLP”) solutions. Cohere creates the necessary prompts that get fed into the model. Cohere also has a prompt engineering lab which checks the responses to the prompts.
In summary, due to these collaborations and shared service models, the EPM teams at Oracle can get the specialized expertise required without growing their team unnecessarily. They do have some data science experts and these folks are regularly leveraged for the innovations being created within Cloud EPM.
What can we expect for the future of Oracle EPM reporting + IPM? Oracle’s Vision for reporting.
One of the most interesting points of the conversation was what Oracle EPM reporting customers can expect in the future. While it doesn’t seem like a lot of Ai innovations are being released in EPM reporting right at this moment, that doesn’t mean there’s a lack of vision or effort happening behind the scenes. As Al commented, “GenAi is the next leap forward with Oracle EPM reporting.”
The Oracle EPM reporting team is vetting ideas with other Product Managers and working on a GenAi roadmap right now. We had to wait to have this interview until March because February was when the team had their Sr. Development Manager meetings.
There are many potential paths. Imagine a scenario where the Cloud EPM solution platform pinpoints why something happened and gives recommendations on how to resolve it. Sounds cool, right? This is within the realm of possibilities.
Want some real insight on where Oracle EPM reporting is going? A Narrative Reporting concept demo was given at Oracle Cloud World 2023. Luckily, it was posted publicly so you can watch it for yourself on YouTube (and it’s short – only a minute long!):
What this is all pointing to is an idea the EPM reporting team calls “touchless reporting.” It’s the vision that Oracle is currently moving towards. In the concept video above, for instance, many of the pieces already exist today, but they just haven’t been extended to EPM reporting yet. The Oracle EPM teams are busily working behind the scenes on how this would all work and which exact reporting tools this would touch.
Side note from Oracle: as with most “concepts” – whether technology, automobiles, or others – user experience and actual implementation may vary from what’s portrayed in the vision demo referenced above. But it is the north star driving the project.
Here are the high-level components Oracle hopes to encompass with EPM reporting in the future:
- Generated narratives that happen directly in reporting
- IPM Insights with causality/explainability generated by AI and enhanced via collaboration passing through to reporting
And they want it all to be easy. Al relayed an example of how this scenario might look. Imagine a customer telling their Cloud EPM solution “give me this past quarter’s North American income statement” and the Cloud platform generating a report package with visuals, tables, and narrative.
The full solution will be rolled out in steps, starting this year. They imagine the first release will improve upon existing functionality, then become more interactive with the consumer as it matures.
When we will be able to see progress on this vision?
Right now, the team has a lot of backlog they need to push through. For instance, the full removal of Financial Reporting has been delayed (for those of you who missed the PSA) to mid-2025. They also have some promised features they need to focus on. Regardless, you’ll start to see the beginnings of the future of EPM reporting later this calendar year.
For instance, with insights today, end users can gather details and numbers. Oracle wants to add in the ability to generate narrative on it. They also want to be able to do causation in reports by giving users the ability to drill down into issues and have the solution make call outs.
They hope the first pieces of causation will come out in the second half of this calendar year (#safeharbor). What they envision for the initial rollout is to go down one level below. Narrative will be generated in conjunction and carried forward for further analysis. Then they’ll build upon that in future versions.
Thanks to Al, Tom, and Dave for sharing your generous time! I’m looking forward to seeing what amazing reporting innovations you come out with!
Additional Resources
Curious about Oracle’s Ai as it exists within EPM reporting? Here are some links to help get you started:
- Narrative Reporting page: https://www.oracle.com/performance-management/narrative-reporting/
- The link to Oracle’s IPM tutorials page: https://docs.oracle.com/en/cloud/saas/planning-budgeting-cloud/tutorials.html
The Full Oracle+AI Blog Series
Did you miss a post? Here is the YTD list of posts in this Oracle + AI blog series:
woman in epm 
3 thoughts on “Oracle + AI Series: Next Gen Oracle EPM Reporting”