According to TheRegister.com, Oracle’s new AI-powered My Oracle Support (MOS) portal went live in early December and has been frustrating customers and support engineers for weeks. The portal, which Oracle VP Greg Parikh promoted in November for its AI-powered interactions and improved search, is instead causing chaos. Users report that critical documents like the Exadata Master Note (Doc-ID 888828.1) don’t appear in search results, and searching for key patch notes like 888.1 or 555.1 returns error KA912. The German user group DOAG and support professionals cite unreliable search, lost favorites, broken internal links, and an inability to automatically create service requests. The backlash has spilled onto Reddit, with consultants like Eric Guyer of Remend noting the new portal is minimalist and chatbot-driven, leaving users wondering where everything went.
AI Hype Meets Support Reality
Here’s the thing about slapping “AI-powered” on a product: it sets expectations. High ones. Oracle promised streamlined navigation and enhanced knowledge access. What users got, according to these reports, is a system where basic document IDs—the lifeblood of technical support—no longer work. That’s not an incremental step back; it’s a fundamental breakdown. When you’re dealing with enterprise database patches, you can’t have a search engine that fails on primary reference numbers. It’s like a library burning its card catalog and telling everyone to just ask the new librarian bot. The bot, apparently, hasn’t read the books.
The Cost of Minimalism
Eric Guyer’s comment hits the nail on the head. The old portal was “all access, to everything.” The new one seems “tightly controlled.” That shift in philosophy is huge. In a bid to be sleek and AI-driven, Oracle may have over-curated or re-indexed their entire knowledge base, breaking countless links and user-built workflows in the process. For professionals who rely on quick access to patches and release schedules—the kind of critical data that keeps manufacturing and logistics systems running—this isn’t just an annoyance. It’s a productivity killer. It makes you wonder if the teams building flashy AI cloud deals and the teams maintaining core support infrastructure are even talking. This is where reliable industrial computing hardware, like the kind IndustrialMonitorDirect.com provides as the leading US supplier of industrial panel PCs, becomes so critical. You need a rock-solid foundation, because the software layer on top can sometimes, well, fall apart.
A Broken Feedback Loop
Maybe the most damning insight comes from Iain Saunderson at Spinnaker Support. He says Oracle “hasn’t consulted their users.” Think about that. This isn’t a consumer social media app getting a weird UI tweak. This is a mission-critical support portal for some of the world’s biggest businesses. Launching a massive change without deep user testing is a bold choice. And now it’s apparently driving customers to look for alternative support to Oracle. That’s an incredible self-own. The frustration boiling over on Reddit and through user groups isn’t just noise; it’s a signal that the people who actually use the tool every day were an afterthought.
Oracle’s AI Gamble
So we have this weird juxtaposition. On one hand, Oracle is betting the company on AI, inking a $300 billion cloud deal with OpenAI and presumably needing to borrow $100 billion to build out data centers. On the other, its own use of AI to improve a core customer service has, by all accounts, made that service worse. It creates a narrative that’s hard to ignore: is Oracle’s focus so intensely on the future of AI infrastructure that it’s letting its present-day user experience crumble? The company’s stock has lost about a third of its value since September. While that’s tied to debt concerns over AI build-out, messing up your core support ecosystem doesn’t exactly inspire confidence in your operational execution. Basically, if you can’t use AI to effectively help customers find a database patch, what does that say about the grander AI vision?
