AI in Oracle EPM and ERP: What’s Real in 2026

Two years ago, we shared an early look at how Oracle was embedding generative AI into its Fusion Cloud Applications. At the time, it was mostly vision demos and roadmap slides. A lot has changed since then — AI in EPM and ERP has moved from concept to production, and finance teams across Australia and New Zealand are starting to use it in their day-to-day work.

Here’s what’s actually landed, what’s working, and what to watch next.

What’s Live in Oracle EPM Cloud Today

Oracle has been shipping AI capabilities through its monthly EPM Cloud updates. The major features that are now generally available include:

Predictive Planning uses time-series forecasting to automatically generate projections based on historical data. Finance teams can run Auto Predict to create a statistical baseline forecast, then compare it against their manually prepared plans. This is particularly useful for rolling forecasts — instead of building every month from scratch, the AI generates a starting position that analysts can refine. Organisations using Predictive Planning are reporting meaningful reductions in forecast preparation time.

IPM Insights (Intelligent Performance Management) analyses planning and consolidation data to detect anomalies, trends, and outliers automatically. Rather than waiting for someone to notice that a particular cost centre is trending 15% above budget, IPM Insights flags it proactively. It’s an early-warning system for finance teams who don’t have time to manually scan every data point.

The Planning Agent is one of Oracle’s most significant recent additions. It’s a conversational AI assistant embedded within EPM Cloud that can generate predictions, explain forecast variances in natural language, run root-cause analysis, and model what-if scenarios. You can ask it “why did travel expenses increase in Q1?” and it will analyse the underlying data and provide a contextual explanation. This is a genuine step-change in how non-technical finance users interact with planning data.

Narrative Reporting with GenAI now includes AI-generated commentary. When building management reports or board packs, the system can draft narrative summaries of financial results based on the underlying data — highlighting key movements, variances, and trends. The output is editable, so skilled finance professionals retain full control over the final content.

What’s Live in Oracle ERP Cloud

AI capabilities in Fusion Cloud ERP have expanded significantly:

Intelligent Document Recognition automates the processing of invoices, receipts, and other financial documents by extracting data from scanned documents and routing them through approval workflows.

AI-Assisted Journal Entries suggests journal entries based on historical patterns, reducing manual data entry for recurring accruals and adjustments.

Project Planning with AI helps project managers build dynamic project plans that optimise resources against financial objectives — the demo we shared in 2024 is now a working feature.

Cash Flow Forecasting uses AI to predict future cash positions based on receivables, payables, and historical cash flow patterns.

What About OneStream?

OneStream has been equally aggressive with AI. Their SensibleAI portfolio now includes:

SensibleAI Forecast generates automated statistical forecasts using machine learning, running natively within the unified OneStream platform. Customers are reporting planning cycles that are significantly faster with improved forecast accuracy.

SensibleAI Agents are embedded AI assistants — the Finance Analyst Agent, Operations Analyst Agent, and Search Analyst Agent — that let users query financial and operational data using natural language. These agents now integrate directly with Microsoft Teams and Excel through OneStream’s expanded Microsoft partnership, meaning finance insights surface where teams already work.

SensibleAI Studio provides a library of 30+ plug-and-play AI routines for anomaly detection, benchmarking, outlier analysis, and more. It’s designed so finance teams can activate AI capabilities without needing data science expertise.

What’s Actually Working in Practice

Across our client base, the AI features that are delivering the most immediate value are:

Predictive Planning for rolling forecasts. The ability to generate a statistical baseline forecast automatically — then have analysts focus their time on the areas where human judgement adds value — is saving significant time in every forecast cycle.

Anomaly detection during close. IPM Insights and SensibleAI’s outlier analysis are catching data issues that would previously have made it through to published reports. This is particularly valuable during the time pressure of month-end close.

Natural-language variance analysis. Both the Oracle Planning Agent and OneStream’s Finance Analyst Agent are genuinely useful for answering ad-hoc questions about financial data without requiring someone to build a report. This is democratising access to finance data in ways that weren’t possible even a year ago.

What to Watch Next

The next wave of AI in EPM will focus on three areas:

Autonomous agents that execute multi-step tasks. Today’s AI agents answer questions. Tomorrow’s will execute complete workflows — preparing a forecast, identifying anomalies, drafting commentary, and routing it for approval — with minimal human intervention. Both Oracle and OneStream are building toward this.

Cross-application intelligence. Oracle’s advantage is the breadth of its cloud stack. When EPM AI can draw insights from ERP transactions, HCM workforce data, and SCM supply chain signals simultaneously, the quality of forecasting and planning improves dramatically.

AI governance and auditability. As AI-generated outputs make their way into statutory reports and board packs, finance teams need audit trails that explain how AI reached its conclusions. Both vendors are investing in explainability and governance frameworks.

How to Get Started

If you’re running Oracle EPM Cloud or OneStream and haven’t activated AI features yet, the good news is that most of these capabilities are included in your existing licence. You don’t need a separate AI procurement — you need a plan to enable and adopt what you already have.

At James & Monroe, we’re helping clients across ANZ and Asia activate AI capabilities in their existing EPM environments. The typical engagement starts with identifying the highest-value use cases (usually predictive forecasting and anomaly detection), configuring the features, and training finance users to work with AI-assisted workflows.

Get in touch: Visit jamesandmonroe.com/contact to start the conversation.