Oracle Cloud EPM AI: A Module-by-Module Guide to Embedded Intelligence
Key Takeaways
- Oracle’s EPM AI strategy operates across three distinct tiers: Agentic Experiences built natively into EPM, Agentic Assistants deployed through Fusion AI Agent Studio, and Interoperability with external AI platforms including ChatGPT, Claude, and OCI Agent Platform.
- The AI stack spans Predictive AI for forecasting and anomaly detection, Generative AI for summarization and narrative, and AI Agents for autonomous action.
- Finance teams already on Oracle Cloud EPM have access to these capabilities now.
For the latest, AI feature release visit Oracle’s readiness updates page.
This blog was last updated: 8/31/2026
From Platform to Intelligent Finance System
Oracle Cloud Enterprise Performance Management has always been positioned as the most complete EPM suite on the market. What has changed significantly over the last three years is the depth of AI embedded throughout it.
The 2016 version of Oracle Cloud EPM had one AI feature. The 2026 version has AI embedded across every module, updated on a monthly release cycle, with an agentic roadmap that extends through 2027. Oracle is moving from AI as insight generation to AI as operational enablement, from features that surface information to agents that execute actions. Understanding where the platform is today, and where it is going, is now a prerequisite for any finance transformation conversation.
Oracle’s Three-Tier AI Architecture
Oracle’s AI strategy for EPM is not a single capability. It is a three-tier architecture, each with a different interaction model, deployment approach, and level of implementation effort.
Each tier has a different effort-to-implement profile. Agentic Experiences are embedded and require minimal consulting effort to activate. Agentic Assistants require AI Agent Studio expertise and prompt engineering, medium effort. Interoperability with external platforms requires agentic expertise and custom development, highest effort.
Agentic Experiences
Intelligent agents native to EPM, operating inside the application. Users interact with planning, financial close, and reporting through natural conversation. Oracle describes three progressive capability levels:
- Conversation-assisted Actions – retrieving data, visualizing trends, explaining variances).
- Guided Finance Workflows – progressively assisting with complex tasks like reconciliations and variance investigation.
- Autonomous Finance Operations – agents that generate scenarios, evaluate outcomes, and recommend actions guided by enterprise policies and financial controls. Current examples include the Planning Agent, the Reporting Agent, the redesigned PCM Agent, and EDM’s Change Management Assistant.
Agentic Assistants
These are designed and managed in Oracle Fusion AI Agent Studio. Business Objects expose common EPM capabilities without requiring finance teams to rebuild business logic from scratch. Over 33 named Agentic Assistants are on Oracle’s roadmap across every EPM module covering users, administrators, and implementers. Examples include the EPM Assistant, PCM Performance Assistant, Consolidation Journals Assistant, Account Reconciliation Assistant, and EDM Request Authoring Assistant.
Interoperability with AI Platforms
REST API-based connections enabling Oracle Cloud EPM to work with external AI platforms ChatGPT, Claude, OCI AI Agent Platform, and others. This tier is flexible but requires custom development and the licensing for external AI platforms is separate from Oracle.
This tier got more concrete in August 2026 with new Insights REST APIs built specifically for agentic AI interactions. They give outside systems programmatic access to insights and summary data across Account Reconciliation, Enterprise Data Management, Enterprise Profitability and Cost Management, Financial Consolidation and Close, FreeForm, Narrative Reporting, Planning, Profitability and Cost Management, and Tax Reporting, which is nearly the entire suite.
Enterprise Planning, The Most Complete AI Stack in the Suite
Oracle describes its Planning product with a specific phrase: reimagine FP&A with the Planning Agent. That is the headline capability for 2026, but the AI story in Planning started a decade earlier and has compounded with every release.
Predictive Planning (December 2016) The original AI investment in Oracle Cloud EPM. Predictive Planning uses time-series modeling to automatically generate statistical projections based on historical data — giving planners a statistically grounded starting point for every forecast cycle rather than a blank template.
Auto Predict (August 2020) Automates the prediction generation process without requiring manual configuration of individual forecast models, bringing statistical forecasting to a broader user population.
IPM Insights (November 2021) Intelligent Performance Management Insights continuously scans financial and operational data to identify anomalies, variances, and trends across scenarios — actuals vs. forecasts vs. predictions — and explains them with context-aware narrative summaries. Insight types include Forecast Variance, Prediction Variance, and Anomaly Detection. GenAI narrative summaries were added in April 2025. In April 2026, Oracle extended GenAI summaries to the Account dimension, meaning anomalies detected at the account level now generate plain-language explanations automatically.
** Three more IPM Insights updates have shipped since: a new Period Movement Variance insight type in July 2026, support for viewing insights directly from forms and dashboards (also July 2026), and sibling-related function support added in August 2026, which also extended Period Movement Variance insights into forms and dashboards.
Bring Your Own ML (November 2021) For organizations with data science capabilities, Oracle Cloud Planning supports custom machine learning model integration, allowing finance teams to bring externally built models into the EPM forecasting framework.
Predictive Cash Forecasting (April 2024) Automated collection of all relevant data streams: cash, accounts receivable, accounts payable, payroll, tax, and external bank data, with predictive algorithms generating a daily or weekly cash forecast.
Advanced Predictions (August 2025) A multivariate AI-driven forecasting capability that allows finance teams to forecast outcomes using multiple business drivers simultaneously not just historical financial data, but revenue drivers, headcount, macroeconomic indicators, and operational metrics in a single prediction model.
Monte Carlo Simulations Built into Oracle Cloud EPM Planning’s scenario modeling framework, Monte Carlo simulations determine the likelihood of various scenarios, giving finance teams statistical confidence in decisions rather than point estimates.
Financial Planning Application 2.0 (Roadmap) Oracle is redesigning Planning for agentic experiences. Financial Planning 2.0 introduces a multi-cube architecture with separate Revenue, Expense, and Financial Statement cubes; dimensions based on Chart of Account segments; seamless integration and drill-through to Fusion Analytical Views; blended forecasting methods including Trend, Driver, Prediction, Smart Drivers, and Direct Input; seeded driver-based models for Revenue and Expense; and IFRS18 financial statements for Management Performance Measures and EBITDA financial statements.
The Planning Agent (Roadmap) Native conversational AI embedded within EPM. The Planning Agent gives finance users conversational access to plans, actuals, and IPM Insights allowing them to drill, explain variances, and surface drivers without navigating forms or reports. It accelerates insight-to-action across the FP&A workflow, moving analysts from static report consumption to interactive financial reasoning. What-if analysis against predictive forecasts allows users to simulate scenarios in conversation. The Planning Agent operates through three built-in assistants: the Data Exploration Assistant, the What-if Prediction Assistant, and the Insight Exploration Assistant.
Financial Consolidation and Close (FCCS) — AI in the Close Process
Oracle was named a Leader again in the 2026 Gartner Magic Quadrant for Financial Close and Consolidation Solutions. The AI capabilities embedded in the finance tool are a significant part of that recognition.
IPM Insights (July 2025) – Oracle extended IPM Insights into Financial Consolidation and Close in 2025, bringing AI-driven anomaly detection, trend identification, and forecast bias analysis directly into the consolidation process. In other words, finance teams can surface outliers, hidden patterns, and missing or volatile data during the close, before the numbers are published.
Consolidation Job Analytics (July 2025) – Uses Generative AI to provide automated analysis of consolidation runs giving controllers and finance directors insight into consolidation performance and processing anomalies through generated summaries rather than manual log review.
Financial Analysis Hub, including Validations Hub (Released August 2026) – What was on the roadmap as a standalone Validations Hub shipped in August as part of a larger Financial Analysis Hub, and the scope is bigger than a validations page. It gives finance teams built-in analytic capability across three seeded financial statements (Balance Sheet, Income Statement, Cash Flow Statement), interactive charts and Key Highlights for comparative analysis, and the ability to define and configure custom ratios beyond the out-of-the-box calculations. The Validation Hub component consolidates validations into one centralized page where they can be enabled or disabled as needed, with AI-generated context on what’s driving each flagged item.
One adoption caveat worth calling out to readers directly: this feature is only available for Dense/Sparse optimized (DSO) applications. Organizations running non-DSO applications don’t have access to it yet.
Financial Consolidation and Close Agentic Assistants (Partially released as of August 2026) Two FCC assistants have shipped since Oracle’s original five-assistant roadmap for this module. A general Financial Consolidation and Close Assistants release went live in July 2026, and a Financial Consolidation and Close Diagnosis Agent followed in August 2026, built specifically to help service administrators diagnose data retrieval performance issues.
Still on the roadmap: the Consolidation Approvals Workflow Assistant, Consolidation Journals Assistant, and Consolidation Rule Assistant.
Account Reconciliation — AI-Suggested Matching and the Reconciliation Agent
Oracle describes Account Reconciliation’s core value clearly: close faster by automating account reconciliations and transaction matching.
Transaction Matching with Intelligent Automation – Oracle’s auto-match engine can match millions of transactions in minutes with one-to-one, many-to-one, and many-to-many matching rules across unlimited data sources. The intelligent automated match feature suggests matches for confirmation or rejection, surfacing AI-generated recommendations rather than requiring manual evaluation.
Account Assignment Assistant (December 2025) – Predictive AI capability that automates the assignment of reconciliations to preparers and reviewers.
Transaction Matching Assistant (December 2025) – Predictive AI capability that assists with complex transaction matching scenarios surfacing suggested matches and supporting human decision-making on exceptions.
Account Reconciliation AI Assistants (June 2026) – The Transaction Matching Assistant for Administrators helps run jobs like Train Match Prediction, Auto Match, Run Auto Alert, and transaction archiving and purging. The Account Reconciliation Assistant for Administrators helps manage periods and run predicted-attribute jobs. The Account Reconciliation Assistant for Users helps with day-to-day reconciliation tasks like viewing details and adding comments. The Account Reconciliation Run Report Assistant for Users runs Reconciliation Compliance and Transaction Matching reports.
Setup requires a Fusion Applications environment with AI Agent Studio, the Service Administrator role in Cloud EPM, federation between the Fusion Applications and Cloud EPM identity domains, and a confidential application in Cloud EPM’s identity and access management.
Still on Oracle’s roadmap for this module: many-to-one and one-to-many match predictions, pre-submission analysis for potential rejection, and NLP filtering for reconciliations.
Profitability and Cost Management — AI Enters Allocation
Oracle’s Profitability and Cost Management Agent automates allocation.
IPM Insights in Enterprise PCM (August 2024) – Anomaly detection and variance analysis in profitability models surfacing unexpected allocation behavior through AI-generated insights.
EPM Assistant for Enterprise PCM (April 2026) Fusion AI Agents integration extended to Enterprise PCM in April 2026 enabling organizations to create custom AI agents that interact with PCM business processes through AI Agent Studio.
PCM Agent: Enhanced and Redesigned (August 2026) – The PCM Agent got more than a version bump. Oracle rebuilt it for a true conversational experience directly inside the application’s key modeling workflows, and embedded it in the screens finance teams actually use day to day: Models, Designer – Waterfall Setup, Designer – Mass Edit, and Calculation Control. The redesigned agent lets users create and modify models and business rules, run calculations, and perform POV operations in natural language without switching to a separate interface.
This comes with a hard cutover date worth flagging to any team currently running PCM. Both the original PCM Agent and the enhanced version are available through the August release, but the original PCM Agent is retired starting the October 2026 (26.10) update. Starting with the November release, the Analyze Performance option on the Calculation Analysis page will only be available through the new PCM Agent. Teams that rely on the current PCM Agent experience have a two-month window to get comfortable with the redesigned one before the old path disappears.
PCM Agentic Assistants (Roadmap) PCM Performance Assistant for Administrators, PCM Interrogation Assistant for Users, and PCM Calculation Assistant for Administrators.
Narrative Reporting – GenAI Today, Reporting Agent Coming
Oracle’s Narrative Reporting has moved from a set of separate GenAI features toward a single, unified conversational experience that goes from answering questions, to investigating issues, to helping users edit and capture report insights.
Generative AI for Management Reporting Narratives (October 2024) – When building management reports or board packs, the system can draft narrative summaries of financial results based on the underlying EPM data highlighting key movements, variances, and trends. The output is editable, preserving finance professional judgment while eliminating hours of manual commentary drafting.
Notes Summarization with GenAI (August 2025) – Automatically summarizes notes entered across a report package giving reviewers and approvers a consolidated view of commentary.
Gen AI Causality Analysis for Formula Segments (June 2026) – Causality Analysis, which identifies which Analytic Dimension members are driving a change in a value, was extended to grid formulas including growth rate and variance percentage formulas. Previously this GenAI capability only covered non-formula cells. It applies across Enterprise Profitability and Cost Management, Financial Consolidation and Close, FreeForm, Narrative Reporting, Planning, and Tax Reporting, which makes it the analytical engine underneath a lot of the narrative and reporting capabilities described elsewhere in this guide.
Auto Text Summary Criteria Manager and Zoom Capabilities (March 2026) Enhanced AI-driven text summary controls and zoom capabilities for narrative content giving report authors finer control over how GenAI-generated summaries are scoped and structured.
Reporting Agent ( July 2026) – The Reporting Agent is an AI-powered experience for analyzing report data and gaining insights using natural-language questions. It answers questions about values, trends, rankings, comparisons, and calculations directly from HTML report previews, and supports conversational follow-up while staying grounded in the current report, grid, and point of view.
This isn’t a Narrative Reporting-only feature. It applies across Enterprise Profitability and Cost Management, Financial Consolidation and Close, FreeForm, Narrative Reporting, Planning, and Tax Reporting. A companion feature, Ask Oracle for Library, also shipped in Narrative Reporting, letting users search library content and find reports and other artifacts through natural-language requests.
Narrative Reporting Assistant (Roadmap) With the core conversational layer now live, Oracle’s remaining roadmap for this area focuses on dynamic investigation and causal analysis (investigating drivers behind report values and variances), generating focused analysis views with ranking and filtering, creating reusable grid outputs, and natural-language editing and guided authoring, including editing grid structures through natural language and broader creation of report objects and insights.
Enterprise Data Management (EDM) — AI Enters Governance
Application Registration Assistant: Generate Properties (December 2025) – Generative AI automates the generation of application properties during the EDM registration process reducing manual configuration when onboarding new applications into the master data governance framework.
Change Management AI Assistant (June 2026) –
Available via the Ask Oracle button in an open view, giving users a generative AI chat experience to query, browse, and change enterprise data. Users can find nodes matching specific criteria, review node details and history, make mass changes to nodes, and ask about the status of requests using natural language. Viewpoint queries and change requests are generated during the conversation and can be edited before submission. The assistant is optional and can be enabled in system settings.
EDM Request Authoring Assistant (Roadmap) An Agentic Assistant for EDM request authoring is targeted for release 26B enabling users to author metadata change requests through natural language. Another feature includes an EDM Email Approval Assistant for Users bringing conversational AI and automated governance workflows into the master data management layer.
FreeForm — Full AI Stack for Custom Planning Models
FreeForm received the full suite of Planning AI capabilities in September 2024 and has continued to receive parity updates. Current AI features in FreeForm include Predictive Planning, Auto Predict, IPM Insights, Bring Your Own ML, Advanced Predictions, and EPM Assistant Fusion AI Agents integration. With the Reporting Agent’s July 2026 release also naming FreeForm as an applicable module, organizations using FreeForm for custom strategic planning models, long-range financial scenarios, or bespoke operational planning frameworks now have access to conversational report analysis as well as the existing AI stack.
Task Manager AI Assistants
Task Manager AI Assistants shipped in June 2026 for Enterprise Profitability and Cost Management, Financial Consolidation and Close, Planning, and Tax Reporting. The Task Manager Assistant for Users helps people review task assignments, summaries, and details, and add commentary. The Task Manager Assistant for Administrators helps create new task schedules from saved templates. The Task Manager Run Report Assistant for Users generates summary, overdue-task, and task-comment reports, prompting for parameters as needed.
Tax Reporting — AI Across the Tax Lifecycle
Tax Reporting continues to receive one of the most comprehensive Agentic Assistant rollouts in Oracle’s 2026 roadmap.
Tax Reporting Assistant (July 2026) – The Tax Reporting Assistant shipped in July 2026, giving Tax Reporting users conversational access through Fusion AI Agent Studio. This requires setup and Generative AI to be enabled in the environment before it can be downloaded and configured.
Remaining Tax Reporting Roadmap – From the Tax RTA Assistant for Users, Tax Journal Assistant and Tax Governance Assistant for Administrators, to the Tax Payments Assistant along with Supplemental Data Assistants for Tax. For organizations with Tax Reporting in scope, this represents a meaningful expansion of AI-assisted workflow from return-to-accrual adjustments through governance, journal management, and payment processing.
Watch the AI World Series on YouTube!
Frequently Asked Questions: Oracle Cloud EPM AI
What is the difference between the Planning Agent and the EPM Assistant?
The Planning Agent is an Agentic Experience a native conversational interface embedded inside Oracle Cloud EPM that finance users interact with directly. It delivers text, charts, insights, and predictions through natural conversation. The EPM Assistant is an Agentic Assistant a template-based artifact downloaded from your EPM environment and deployed through Fusion AI Agent Studio to create governed, composable AI workflows. Both are valid approaches with different deployment requirements and use cases.
What is Oracle’s Reporting Agent?
The Reporting Agent is an AI-powered, conversational experience released in July 2026 that lets users ask natural-language questions about report data, values, trends, rankings, comparisons, and calculations directly from an HTML report preview. It supports follow-up questions and stays grounded in the current report, grid, and point of view. It spans six modules: Enterprise Profitability and Cost Management, Financial Consolidation and Close, FreeForm, Narrative Reporting, Planning, and Tax Reporting.
Do I need to enable anything before I can use EPM Assistants?
Yes. Starting with the July 2026 release, Generative AI has to be enabled in your Oracle Cloud EPM environment before EPM Assistants can be downloaded at all. This is a genuine prerequisite, not an optional setting, so it’s worth confirming with your administrator before planning a rollout timeline for any assistant.
Are these AI features available to all Oracle Cloud EPM customers?
Most features are included in the standard Oracle Cloud EPM subscription and are delivered through the monthly update cycle. Some GenAI capabilities depend on data center availability and regional rollout. The EPM Assistant and Fusion AI Agents integration require a Fusion Applications environment with AI Agent Studio enabled, and as of July 2026, Generative AI must specifically be enabled in the environment before any EPM Assistant can be downloaded. EPMI can assess your specific environment and confirm which capabilities are available and active.
Can Oracle Cloud EPM work with external AI platforms like ChatGPT or Claude?
Yes. Oracle’s third tier of AI strategy, Interoperability with AI Platforms, uses REST APIs to connect Oracle Cloud EPM with external AI platforms including ChatGPT, Claude, and OCI AI Agent Platform. This approach is flexible but requires custom development, and licensing for external AI platforms is separate from Oracle. As of August 2026, new Insights REST APIs give agentic AI systems programmatic access to insights and summary data across nearly every EPM module, making this tier considerably easier to build against than it was a year ago. Smart View CoPilot Agent also enables conversational chart building in Excel through this interoperability framework.
How does EPMI help organizations activate Oracle Cloud EPM’s AI capabilities?
EPMI’s implementation practice and managed services team have deep experience with Oracle Cloud EPM’s AI feature set across all three tiers, and we track Oracle’s monthly release cadence closely so our clients aren’t caught off guard by changes like the PCM Agent’s October 2026 deprecation date. For new implementations, we build AI activation into the implementation roadmap, not as a Phase 3 aspiration, but as a go-live deliverable. For existing Oracle Cloud EPM customers, we offer an AI Readiness Assessment that evaluates which features are enabled, which are configured, and where the highest-value opportunities are. Schedule a conversation with our team to start there.