What replaces Agents and iBots?

QuickSight scheduled reports and threshold alerts can cover many alerting and delivery requirements. Paginated report schedules can support scheduled distribution of formatted reports, while recipient-specific requirements should be redesigned around QuickSight security and scheduling capabilities.

25 Sep 2026

Oracle BI to AWS QuickSight: Migration Strategy, Challenges & a Proven Path

Q: Does QuickSight have an equivalent of the OBIEE RPD?

No single equivalent exists. QuickSight's governance is built from shared, certified datasets with calculated fields rather than a three-layer semantic model. Dataset modeling and rationalization therefore become important steps in an Oracle BI to QuickSight migration.

Q: Can QuickSight replace BI Publisher pixel-perfect reports?

Yes. Highly formatted, printable output can be rebuilt as QuickSight paginated reports, while interactive content can be rebuilt as QuickSight analyses and dashboards.

Q: Should we use SPICE or direct query for Oracle data?

The choice depends on data volume, performance requirements, and refresh SLAs. SPICE provides in-memory query performance, while direct query is appropriate when queries need to access the underlying source data directly. Oracle data can also be landed in AWS services such as Amazon Redshift or Amazon S3 with Athena.

Q: How is OBIEE data-level security handled in QuickSight?

OBIEE data-level security is re-implemented using QuickSight security capabilities such as row-level and column-level security. Users and groups can be managed through the organization's identity model, and security rules should be tested against expected results before cutover.

Q: Do we have to migrate the database at the same time?

Not necessarily. QuickSight can connect to supported Oracle sources through a VPC connection, allowing reporting to migrate before the underlying database. Data can subsequently be landed on AWS depending on the target architecture, performance requirements, and cost objectives.

Quick Summary

Oracle BI to AWS QuickSight migration is a re-platforming, not a file conversion: OBIEE's RPD, analyses, dashboards, iBots, and BI Publisher reports each map to a different QuickSight construct. QuickSight has no RPD equivalent, so governed logic is rebuilt as shared, certified datasets first. Figure-for-figure validation against OBIEE remains the gate before cutover.

Amazon QuickSight is serverless, scales automatically, and prices by usage, with generative BI layered on top through Amazon Q. The catch: QuickSight is architected very differently from Oracle BI. No direct equivalent exists for the RPD's three-layer semantic model, and OBIEE's analyses, dashboards, iBots, and BI Publisher reports each map to a different QuickSight construct. This guide covers why organizations migrate, what actually has to move, the challenges by component, and a phased methodology that de-risks the work, the same rebuild-by-intent discipline covered in our Cognos to AWS QuickSight migration guide for a different source platform on the same target.

oracle-bi-publisher-to-amazon-quicksight

Key Takeaways

This is a re-platforming, not a conversion. The RPD, analyses, dashboards, agents, and BI Publisher reports each map to a different QuickSight construct.
QuickSight has no RPD equivalent. You rebuild governed logic as shared, certified datasets with calculated fields, so modeling and rationalization come first.
Pixel-perfect BI Publisher output moves to QuickSight paginated reports; interactive analyses and dashboards move to QuickSight analyses and dashboards.
SPICE vs. direct query is the key performance decision, sized against refresh SLAs. On-prem Oracle data connects through a VPC, or is landed in Redshift / S3 + Athena.
Amazon Q in QuickSight adds generative BI, natural-language Q&A, executive summaries, and data stories, but figure-for-figure validation against OBIEE remains non-negotiable.

Why Organizations Migrate Oracle BI to Amazon QuickSight

The drivers behind Oracle BI modernization accumulate until staying put costs more than moving:

Serverless, no infrastructure. QuickSight removes the WebLogic servers, RPD administration, and capacity planning that OBIEE and OAS demand; it scales automatically to tens of thousands of users.
Cost and licensing. Usage-based, per-session reader pricing typically lowers total cost versus Oracle BI licensing and support, especially where AWS spend already exists.
AWS-native consolidation. Tight integration with Amazon Redshift, Athena, S3, Aurora/RDS, Lake Formation, and IAM Identity Center unifies analytics on one stack, the same AWS-native consolidation covered in our Cognos to AWS QuickSight migration guide and our ODI (Oracle Data Integrator) ETL guide for a related AWS-and-Oracle integration pattern.
Generative BI. Amazon Q in QuickSight brings natural-language Q&A, AI executive summaries, and data stories, capabilities OBIEE never offered, the same generative-BI positioning covered in our Unique Features of Amazon QuickSight overview.
Embedding and reach. Serverless embedded analytics and multi-tenant namespaces make it easier to put dashboards in front of internal and external users.

What Amazon QuickSight Brings to the Table

Knowing the target's building blocks makes the mapping clearer, the same platform-fundamentals-first approach covered in our Unique Features of Amazon QuickSight guide:

SPICE. An in-memory, columnar engine for fast queries, with scheduled refresh; direct query is the alternative for live data.
Datasets, analyses, and dashboards. Datasets hold joins and calculated fields; analyses are the authoring surface; dashboards are the published, shared output.
Paginated reports. Pixel-perfect, printable output with scheduled delivery and row-level security.
Row-level and column-level security (RLS/CLS). Dataset rules enforced across every analysis and dashboard, with users and groups organized in namespaces.
Amazon Q in QuickSight. Generative BI for natural-language questions, summaries, and data stories.

What Actually Has to Move

Reports are the visible tip of the iceberg. A complete Oracle BI migration inventory spans seven areas; skip any one, and the project runs over budget. The same seven-layer inventory discipline covered in our ETL Solutions overview:

RPD semantic model. Physical, business, and presentation layers with joins, hierarchies, logical columns, and aggregation rules.
Answers / ad-hoc analyses. Built on subject areas and Oracle logical SQL.
Interactive dashboards & prompts. Prompts, drilldowns, and guided navigation.
Agents / iBots. Conditional alerts and scheduled bursting and delivery.
BI Publisher reports. Pixel-perfect, printable statements and invoices, the same pixel-perfect fidelity challenge covered in our Automated Migration from BRIO/SQR to OAS guide for a related legacy Oracle reporting technology.
Security. Application roles and data-level security.
Data sources & platform. The Oracle schemas and physical SQL the reports run against, the same source-platform assessment covered in our data lake work.

Common Challenges by Oracle BI Component, and How to Solve Them

Each part of the Oracle BI stack presents its own challenge migrating to QuickSight, the same component-by-component mapping discipline covered in our Tableau to Amazon QuickSight guide for a different source platform on the same target. The table below maps every component to a concrete solution.

Oracle BI Component	Common Migration Challenge	How to Solve It in QuickSight
RPD semantic model	The three-layer RPD centralizes joins, hierarchies, logical columns, and aggregation rules; QuickSight has no equivalent semantic layer.	Rebuild governed logic as shared, certified datasets: model joins in the dataset, and recreate logical columns and aggregations as calculated fields for every analysis.
Answers (ad-hoc analyses)	Analyses are built on presentation-layer subject areas and Oracle logical SQL.	Recreate as QuickSight analyses on the shared datasets; convert logical SQL and column formulas to calculated fields, validated against source.
Interactive dashboards & prompts	Dashboard prompts, drilldowns, and guided navigation don't map one-to-one.	Rebuild dashboards in QuickSight; convert prompts to parameters and controls, and drilldowns to field hierarchies and dashboard actions.
Agents / iBots	iBots handle conditional alerts and scheduled bursting and delivery.	Replace with QuickSight scheduled email reports, threshold alerts on visuals, and, for burst-by-recipient, paginated report schedules driven by RLS.
BI Publisher	Highly formatted, printable statements and invoices with precise, fixed layout.	Rebuild as QuickSight paginated reports for pixel-perfect, printable output with scheduled delivery and row-level security.
Security (roles & data-level)	OBIEE applies data-level security through RPD filters and WebLogic application roles.	Re-implement as QuickSight RLS (dataset rules) and CLS, with users and groups in namespaces via IAM Identity Center; test each rule against expected data.
Data sources & platform	Reports run against on-prem Oracle schemas and hand-tuned physical SQL.	Connect QuickSight to Oracle over a VPC for direct query, or, preferably, land data in Redshift / S3 + Athena and point SPICE datasets at the modern store.

A Proven, Phased Migration Methodology

The reliable path is sequential and evidence-led. Each phase produces an artifact the next phase depends on, which keeps a large migration predictable, the same phased discipline covered in our guide to BI automation for report migration.

Discovery & assessment. Inventory the catalog, RPD subject areas, analyses, dashboards, iBots, BI Publisher reports, and security, with usage statistics and complexity scores to size the effort accurately- the same complexity-scoring approach covered in our key checklist for BI modernization.
Rationalization. Retire unused content, merge duplicates, and classify what remains as interactive versus paginated. This is the single biggest lever on cost.
Data platform & dataset modeling. Land data on AWS where appropriate (Redshift / S3 + Athena) and rebuild the RPD's governed logic as shared, certified QuickSight datasets with calculated fields, the same data-lake-first landing pattern covered in our data lake work.
Analysis, dashboard & report rebuild. Rebuild analyses and dashboards on the shared datasets, and pixel-perfect content as paginated reports, with consistent theming.
Security, delivery & deployment. Implement RLS/CLS and namespaces, configure schedules and alerts, set up embedding where needed, and promote through dev/test/prod.
Validation & parallel run. Compare QuickSight output to OBIEE figure-for-figure, and run both in parallel until numbers reconcile and users sign off, the same figure-for-figure discipline covered in our guide to automating ETL testing with Python.
Cutover, adoption & decommission. Train users, enable Amazon Q, switch delivery to QuickSight, and decommission OBIEE/WebLogic to realize the savings that justified the project.

The Technical Detail That Decides Success

From the RPD to governed datasets, the real deliverable. OBIEE centralized business definitions in the RPD so that a metric meant the same thing everywhere. QuickSight has no three-layer semantic model, so governance shifts to shared, certified datasets with reusable calculated fields. Pushing logic into individual analyses is faster at first but recreates the inconsistency you're migrating away from, the same governed-dataset-first principle covered in our Tableau to Amazon QuickSight guide for a different source platform on the same target.

Logical SQL and column formulas become calculated fields. Oracle logical SQL, EVALUATE functions, and column-level formulas rarely convert one-to-one to QuickSight's calculated-field functions. Filter behavior and aggregation differ, so every converted field should be validated against OBIEE output before it's trusted, the same conversion-and-validate discipline covered in our guide to automating ETL testing with Python.

BI Publisher maps to paginated reports. Interactive QuickSight dashboards are built for exploration, not for a fixed, multi-page printable statement. QuickSight paginated reports cover exactly that operational, pixel-perfect content, the natural home for what BI Publisher handled, with scheduled delivery and RLS.

Data-level security becomes RLS, CLS, and namespaces. OBIEE's data-level security and application roles are re-expressed as QuickSight row-level and column-level security rules, with users and groups organized into namespaces through IAM Identity Center. Legacy security must be re-implemented and tested per rule; it does not carry over automatically.

SPICE or direct query, decide per dataset. SPICE gives fast, scalable in-memory performance but refreshes on a schedule; direct query keeps data live but pushes load to the source and forgoes SPICE speed. Choose per dataset against real refresh SLAs, and provision VPC connectivity for any on-prem Oracle source, the same SPICE-vs-live-query decision covered in our Tableau to Amazon QuickSight guide.

Best Practices for a Low-Risk QuickSight Migration

Model datasets once, centrally; certify them, and let analyses consume them.
Decide SPICE vs. direct query per dataset, against real refresh SLAs.
Rationalize before you rebuild; never migrate content nobody opens- the same rationalize-first discipline covered in our key checklist for BI modernization.
Land data on AWS where possible rather than relying long-term on direct query to on-prem Oracle.
Validate every figure against OBIEE before decommissioning anything, the same reconciliation discipline covered in our guide to BI automation for report migration.
Run OBIEE and QuickSight in parallel through at least one full reporting cycle.
Plan Amazon Q enablement and reader adoption as part of the project, not after it.

Frequently Asked Questions

Does QuickSight have an equivalent of the OBIEE RPD?

No single equivalent. QuickSight's governance is built from shared, certified datasets with calculated fields rather than a three-layer semantic model, which is why dataset modeling and rationalization come first in a migration, the same governance-rebuild discipline covered in our OBIEE to Microsoft Fabric guide for a different target platform.

Can QuickSight replace BI Publisher pixel-perfect reports?

Yes. Highly formatted, printable output moves to QuickSight paginated reports, which support precise layout, scheduled delivery, and row-level security, while interactive content moves to QuickSight analyses and dashboards.

Should we use SPICE or direct query for Oracle data?

It depends on data volume and refresh SLAs. SPICE gives the best performance for most dashboards; near-real-time needs favor direct query. Landing data in Redshift or S3 + Athena generally gives the best SPICE experience.

How is OBIEE data-level security handled in QuickSight?

Through row-level and column-level security dataset rules, with users and groups organized in namespaces via IAM Identity Center. Legacy security is re-implemented and tested per rule; it does not migrate automatically.

Do we have to migrate the database at the same time?

Not necessarily. QuickSight can direct-query on-prem Oracle over a VPC, so reporting can migrate first. Landing data on AWS (Redshift / S3 + Athena) is recommended as a follow-on for performance and cost, the same phased-landing approach covered in our data lake work.

What replaces Agents / iBots?

QuickSight scheduled email reports and threshold alerts cover most alerting and delivery, while paginated report schedules driven by RLS handle burst-by-recipient distribution.

Planning an Oracle BI to QuickSight Migration?

DataTerrain is an AWS partner with 17+ years in Automated BI migration, from OBIEE, Oracle Analytics, and BI Publisher to Amazon QuickSight, with automated assessment, governed dataset modeling, and figure-for-figure validation- the same broad platform coverage reflected in our best data analytics services and reports conversion services.

Talk to our migration team