Interesting topic, Siddharth. I think Oracle should support self-hosted LLMs, as they could be a good option for organizations that need more control over data, security, and customization.

Absolutely agree. Scaling Enterprise AI is less about the model itself and more about having reliable data, strong governance, and clear business context. I think when AI is integrated into real workflows, that’s where it can deliver sustainable value.

You may not need separate custom models for each invoice format. The built-in Document Understanding model can handle different layouts, so I’d suggest testing it first. If the extraction isn’t accurate enough, you can then consider training a custom model based on your invoice variations.

The move toward a more streamlined Workbook authoring experience looks promising. Separating presentation configuration from the core editing experience should make the workflow cleaner, while still keeping the key capabilities easily accessible. The new navigation approach may take some adjustment for existing users, but…

Great to see Oracle continuing to invest in AI, analytics, and data platform innovation. The roadmap provides valuable insight into what's coming next and helps organizations plan their analytics strategy more effectively.

DV uploaded datasets are typically not migrated through catalog archive/unarchive. If you're using uploaded Excel/CSV files, they may need to be migrated separately or re-uploaded. You may also want to check whether your OAS version supports BAR snapshot migration, which can help preserve more artifacts during migration.…

The invoice payment date is generally not available by default in the PPM - Project Costs or Project Invoices subject areas. This information is typically sourced from the Payables module. To include the invoice payment date in your report, you may need to create a custom OTBI subject area extension, use a BI Publisher…

I have resolved a similar issue while migrating BI Publisher content. In my case, the .xdo, .xpt, .xsldata, and .cfg files were not getting archived, while DV and Analyses were migrated successfully. I verified the catalog permissions, archived the parent BI Publisher folder using Catalog Manager (Offline Mode), and…

The invalid_request error usually indicates that the request payload does not match the AI Agent's expected input schema. Verify the JSON body, parameter names, data types, authentication token, and required headers (Content-Type: application/json). If possible, share the Postman request body (with sensitive data masked)…

Hi Shrikrushna, I faced a similar issue before. It was caused by an Office and BI Publisher version mismatch. Try running the installer as Administrator, ensure Office and BI Publisher are both 32-bit or 64-bit, and repair Office before reinstalling. Hope this helps!