Oracle

Oracle Fusion HCM Analytics gives ready-made HR reports to companies that run Oracle Fusion Cloud HCM. First, settle who sees which records. Then check how much payroll and salary history it keeps. Also check what happens to typed AI questions. Oracle licenses it by employee, with extra charges for some storage, compute, outside data and AI use. After the contract ends, you get 60 days to take your content out.

By Ashish Manchanda Published 10 Nov 2026 Last updated 10 Nov 2026

Oracle’s analytics for Oracle HR data, licensed by employee

In one line

Oracle’s analytics product for Oracle Fusion Cloud HCM turns its data into ready-made HR reports. Oracle manages the link that copies the data, and the data model. AI features sit on top.

Buy it if

You run Oracle Fusion Cloud HCM and want reports Oracle has already built. You are happy to plan the set-up with an Oracle partner, which Oracle recommends. Someone on your side can own the access rules.

Do not buy it if

You cannot yet say who may see staff records. Oracle can limit access to the people a role looks after. Some areas, such as Talent Acquisition, Succession Management and Diversity Analysis, sit outside that limit. Do not sign before you hold a price per employee in writing.

Price

Oracle licenses HCM Analytics by employee. Oracle offers to let you get started for free. Extra storage, extra compute and large volumes of outside data cost more.

Oracle copies your HR data into its own store and builds the reports

Before you judge the AI, look at how Oracle builds the reports underneath. That tells you how much work the product could save your team.

Think of the product as a second copy of your HR data, built for reporting. Oracle copies data from your Oracle Fusion Cloud Applications. It loads the copy into a separate Oracle data store. Oracle calls that store an Autonomous AI Lakehouse. Business users then build dashboards in Oracle Analytics Cloud. Oracle manages the pipeline, which is the link that copies the data across. It also manages the data model and the ready-made content. Oracle’s set-up team includes a data engineer or model administrator for customisation. HR teams can also extend the data model without coding.

Data refreshes on a schedule. It can run daily, weekly or monthly. You can add extra refreshes during the day.

The ready-made content covers headcount, hires, transfers, turnover and time to hire. It also covers span of control, which is how many people each manager leads. Compa ratio compares a person’s pay with the middle of their pay range. Oracle tracks diversity indicators against goals. It splits them by gender, pay grade and ethnicity. It also splits them by tenure band, age band and country. Diversity Analysis tests selection, promotion, salary and termination rates across demographic groups.

People Leader Workbench gives people leaders AI recommendations. Examples are whether to promote someone, suspend requisitions or reinvest in talent. A requisition is a request to fill a job. Profile Matching scores each worker against job profiles. Each worker gets an overall score and a skill gap score. It also suggests the upskilling a worker needs to become eligible for specific job profiles.

Data from outside Oracle can join the same store. Two examples are Salesforce and E-Business Suite. You can also create a Fusion Data Intelligence set-up for outside data. That works even with no Oracle Fusion Cloud Applications product.

The AI Assistant lets business users ask questions in everyday words. It turns each question into a query against the data model. It returns charts, summaries or narrative. It can use agents to find trends, recommended plans of action and areas of improvement.

Not every feature is on general release. On 27 October 2026 Oracle announced that Data Agents are generally available. Fusion Operational Analytics is available on a limited basis. It is an add-on. It is not available on standalone instances. Deep Insight Agents are planned for general availability within the next 12 months. Oracle labels the Taleo and Snowflake connectors, the MCP server and Data Share as Preview. Preview means Oracle has not yet made the feature generally available.

Our guide to people analytics software covers the rest of the market.

Oracle gives different counts of ready-made measures and connectors

Buyers compare how many ready-made measures and connectors a product has. Oracle’s own figures for both differ, so check which count applies to you.

Oracle’s own pages give different counts for the ready-made measures. Its product page puts the number of best practice HCM KPIs at more than 1,000. Its comparison table puts it at more than 2,000 ready-to-use best practice KPIs, metrics, dashboards and reports. Oracle’s implementation guide counts over 2,000 prebuilt key metrics. Those span HR, finance, supply chain and customer experience. An undated Oracle ebook for HCM Analytics puts it at more than 100 prebuilt KPIs. Ask Oracle for the list of measures you would get.

The connector count differs too. Oracle’s product page puts it at more than 50 native self-service connectors. Oracle’s administration guide puts it at more than 30. Ask which ones your own systems need.

Oracle runs the data feed, and access follows your Oracle HR rules

These are the parts of Oracle’s offer that work in your favour. They cover who runs the plumbing, who sees what and how the AI treats your data.

Oracle runs the data feed and updates

Oracle manages the data pipeline, the data model and the ready-made content. That spares your team from keeping the pipeline running. Oracle also applies platform releases by itself. Oracle publishes a 99.9% service-level objective for Fusion Data Intelligence. That is its target for how often the service is available. Where your data sits follows the region you order. Oracle offers commercial, government and EU sovereign regions. They include Frankfurt, London, Ashburn, Hyderabad and Mumbai. Oracle stores your content in the region for the services you order. It may replicate the content within that region.

Access reuses your Oracle HR rules

A line manager’s access follows the reporting lines already set in Oracle Cloud HCM. So you do not rebuild the reporting lines for analytics. Oracle also offers access by area of responsibility. That limits a user to the people their role looks after. It matches the profile in Oracle Cloud HCM. In Oracle’s onboarding plan, it is the recommended option for HR analysts. Oracle can also limit users to the values assigned to them, such as a department or country. The AI Assistant works within each user’s own permissions.

Diversity checks skip groups under 5 employees

That keeps a result from resting on a handful of people. Diversity Analysis checks hires, exits, promotions and pay by gender, ethnicity and age. It runs every day on current data. It gives statistical evidence of potential bias. It makes no predictions, because it analyses existing data.

Oracle Analytics’ AI does not train on your data

It uses foundation models from Oracle Cloud Infrastructure (OCI) Generative AI. Oracle does not use customer data to train those models. It uses them without fine-tuning. Admins can turn off selected AI features. They can switch to an older model that incurs no generative AI charges. Oracle has zero data retention agreements with Google and xAI for models used through OCI Generative AI.

It suits companies that already run Oracle HR, often with Oracle finance

The product suits companies that already run Oracle Fusion Cloud HCM, especially alongside Oracle finance. One shared data model covers HCM, ERP, SCM and CX. So HR and finance questions use the same data.

Heathrow Airport chose it for finance and HR together

Heathrow picked Oracle Fusion Data Intelligence for ERP and HCM. Its finance and HR groups now use it as their strategic analytics solution. Heathrow serves almost 85 million passengers a year. It employs over 90,000 people on site.

Oracle reports faster payroll work at Baylor University

Oracle reports Baylor paired HCM Analytics with Oracle Payroll Activity Center. Payroll data checks became 65% more efficient. Payroll processing time fell by 49%. Those results belong to the two products together, not to HCM Analytics alone. Separately, Oracle reports that joining Baylor’s finance and HR apps made one analysis easier. It is the financial impact of HR policies for more than 4,000 faculty and staff.

Payroll Balances is in use at 90,000 employees

Oracle reports customers with 90,000 employees have implemented it. Payroll Balances is the part that reports payroll balances for each person. So it has run at very large employers.

Who sees what, how much history it keeps and how the AI treats questions

These limits matter most to whoever owns access to employee data. They cover who sees which records and how much history the store keeps. They also cover what the AI does with typed questions.

Some roles see all HCM data

A user with the HR Analyst role sees all HCM data with no restrictions. Decide who gets that role before launch. HR, HR Analyst and Analyst users make up around 0.3% to 3% of the workforce in Oracle’s onboarding plan. Line managers, at up to 10% of the workforce, get access through their reporting line. Some areas also sit outside access by area of responsibility. Examples are Talent Acquisition, Succession Management and Diversity Analysis.

Own-data access is changing

A setting called View Own Context will default to No for new customers from release 26.R3P3. Then no user gets default access to their own data. Under the current default, all users see their own records. Ask which default your set-up will start with.

History has limits

Oracle documents two known issues with history. Payroll Balances or Payroll Costing data can be kept for three years at most. Data older than three years is lost after a full load or what Oracle calls a TDM upgrade. The second issue is about pay. In updates that load only changes, a change to some pay measures overwrites the earlier value. Oracle calls them compensation zone measures. That history is lost. Decide what history you must keep in your own records.

Fixing a wrong load uses limited full reloads

Oracle allows one on-demand full reload per source each month. That reload includes a reset of the data warehouse for all sources. Oracle also allows four scheduled or on-demand full reloads per week per source. They apply to one area of reports or to your own added data. Going past the limits means contacting Oracle Customer Support. The limits are subject to change. Oracle recommends running a check called HCM Diagnostics in Oracle Cloud HCM before each data load. If problems remain, the load rejects those records.

Typed AI questions go to the model as written

Oracle recommends keeping personal details out of prompts. Requests may be routed to another OCI region. The model may generate responses based on patterns in data rather than factual accuracy. Admins can let the Assistant use information from outside your own data. A second tool, AI Concierge, answers questions about how your Oracle set-up is arranged. It is on by default in every customer set-up. It acts only with the admin’s consent.

Connecting your own AI tool carries Oracle’s warning

Oracle calls the connection an MCP server and labels it Preview. Oracle documents connecting it to Claude Desktop. The HCM tools cover workforce, payroll, compensation, attrition, skills and forecast analysis. Oracle warns that the MCP server introduces security risks. Users should only have access to data that can be shared with large language models. The tools run with the permissions of the person who connects.

It can point to individuals who may leave

It can analyse performance by individual, team or organisation. The aim is to find employees who may be a potential flight risk. One Oracle metric counts workers rated high risk of loss in a talent review. It counts those who terminated within one year of that review. Oracle also offers to predict which employee groups are likely to leave. Decide before launch whether managers may see a risk of leaving for a named person.

Security reports may vary

You need a report that covers this service in your country. Oracle may run independent reviews such as SOC 1 or SOC 2 reports. The scope may vary by service and country. Oracle reports attestations for each line of business. Customers can get more from their sales representative.

Oracle licenses it by employee, and you get 60 days to take your data out

The bill has more parts than the licence fee. Here is what costs extra, and what you can take with you when the contract ends.

Oracle licenses HCM Analytics by employee, so your headcount drives the bill. Oracle calls this the employee metric. Oracle offers to let you get started with Fusion HCM Analytics for free. Get the price per employee in writing, and ask what it covers.

Where the data goes changes the licence you need. To use the data with third-party tools or your own applications, you need the employee metric. You need it to export the data to a third-party store too. Exporting to external storage requires enough user licences for access through the external system.

Some data volumes cost extra. Oracle loads data from Oracle Fusion Cloud Applications through ready-to-use pipelines. That includes HCM data. Data from Taleo, PeopleSoft and E-Business Suite loads the same way. That data is included in the subscription. The 50GB limit does not apply to it. Other external data is limited to 50GB, and Oracle charges for more. Data your own team loads in its own custom way is not included in the subscription. Oracle calls that custom ETL into a custom schema. It is subject to the 50GB limit. Salesforce data is included only when you buy it with the Oracle Fusion CX Analytics subscription. Scaling up compute and storage incurs additional usage charges. Scaling up the Lakehouse is billed at the prices in your Oracle Cloud Infrastructure rate card.

AI use is metered too. Oracle meters AI use in Oracle Analytics Cloud. It uses its own units, called AIDP Units. Customers over the thresholds may buy more units. They may also opt in to usage-based billing or switch to a lower cost or free model. Charges for going over start on or after 30 September 2026.

Leaving has a clock on it. For 60 days after termination, Oracle makes your content available in a structured, machine-readable format. Oracle either sends it by a secure method or keeps the service open for you to collect it. For free trials, Oracle has no obligation to keep content after that period. Ending an active instance deletes the Oracle Analytics Cloud and Lakehouse instances. It deletes their data and customisations too. Identity users and groups are not deleted.

Your own team can reach the stored copy while the contract runs. You can reach the Lakehouse directly in three ways. They are an Oracle Analytics Cloud connector, a web tool called Database Actions and a desktop program called SQL Developer. Data Share is a Preview feature that copies processed data to an outside store. It writes table files to OCI Object Storage, which is Oracle’s cloud file storage. It has licensing restrictions and supports only public instances.

The company behind the contract matters too. Oracle’s own headcount has fallen, so get your support contacts written into the order form. Oracle is a publicly traded company under the symbol ORCL. RTE, citing CNBC, reported on 1 April 2026 that Oracle was laying off thousands of employees. BusinessToday reported on 23 June 2026 that Oracle had 141,000 employees at 31 May 2026. That was down from around 162,000.

You could also get these reports and answers from agents built on your own systems. We have mapped what that looks like. The HR Agent Map

Ten questions for Oracle before you sign

Put these to Oracle before you sign. Each one comes with a good answer to listen for.

What is the price per employee, and what does it cover? Good answer: A written price that also covers storage, compute, outside data and AI use.
Which of our roles will see Talent Acquisition, Succession Management and Diversity Analysis data? Good answer: A written map of roles against subject areas. Each area outside access by area of responsibility has named people.
How many years of payroll and salary history do we keep? What survives a full reload? Good answer: A stated length for each type of data, and a way to store older data ourselves.
How many full reloads can we run each month, and what if we need more? Good answer: A contract limit that is enough to fix a data error without a support ticket.
Which AI features are on at launch, and where do AI requests run? Good answer: A list of features on at launch, a named admin for each and a named region.
Which of the features we want are generally available today? Good answer: A written stage for each feature, with a date for anything still in Preview or limited release.
Which security report covers Fusion Data Intelligence in our region? Good answer: A named, dated report that covers this service, not another Oracle line of business.
Is the 99.9% service-level objective in our contract? Good answer: Yes, as a written commitment with a stated remedy if Oracle misses it.
What can we take out in the 60 days after the contract ends, and in what format? Good answer: A file list and format agreed in the contract, and a deletion date for what is left.
Which ready-made measures and connectors will we actually use? Good answer: A named list of both, tested on our own data before we sign.

Alternatives to Oracle

The alternatives are Workday, SAP, One Model, Visier, SplashBI, Crunchr and buying nothing new.

Workday and SAP sell the same kind of product from the maker of an HR system. Workday People Analytics analyses Workday HCM and talent data. SAP People Intelligence is an add-on licence to SAP Business Data Cloud. It connects to SAP SuccessFactors HR data and beyond. If you run Workday or SuccessFactors, look at those two first.

One Model, Visier, SplashBI and Crunchr connect to Oracle from outside. One Model extracts data from Oracle HCM Cloud through REST APIs on a schedule you choose. It also brings together HRIS, ATS, payroll, engagement, performance and finance data. It adds other business data too. All of it sits in one HR data model. Visier has an Oracle Fusion connector among its HR system connectors. SplashBI has a native Oracle HCM connector. Crunchr connects to Oracle Fusion HCM among its HR systems. If your HR data sits in more than one system, compare how each tool joins them. Ask who owns the stored copy.

Buying nothing new is the last route. Keep the reports you have until you hold a list of the specific measures you would gain. Our comparison of 33 people analytics products sets the options side by side.

More in this guide

The People analytics guide
Compare the 33 researched products what each brings together, what its AI does, what it costs and the contract terms.
Mid-market companies which people analytics products suit 200 to 2,000 people, what they cost and what to ask.
Enterprise companies which people analytics products suit 2,000 people or more, what they cost, how contracts run and what to ask.
Companies whose HR system has analytics what the analytics in your HR system covers, what it costs and when to buy something separate.
Companies using their own AI tool or data warehouse how each product lets your own AI tool or data warehouse use the data, what it costs and what to check.