New Oracle Analytics AI Assistant available! Getting started in 3 steps:
https://blogs.oracle.com/analytics/post/oracle-analytics-ai-assistant-getting-started-in-3-steps
New blog from Product Marketing by @Barry Mostert-Oracle :
https://blogs.oracle.com/analytics/post/ai-powered-experiences-for-analysts-and-consumers
Ask your questions to the product management in this thread.
Hi - love this, saw it from your LinkedIn post! I created a dataset in the environment and went to inspect and still see only the old menu of options. I do not see a drop down for "Assistant" to Train it or the "Assistant tab as referenced in the blog. Is this not yet available in FDI? We are on 24.R3. Thanks!
Hi - this is a very nice feature. I saw this from your LinkedIn post! I went to inspect on the dataset in FDI environment and only see the old menu options. I do not see a drop down for Assitant to Train it or Assistant tab as mentioned in the blog. We are on release 24.R3 on FDI. Thanks!
Glad everyone is excited about this feature.
It is being rolled out progressively by region and shape. Also, note that that the initial release supports datasets with Subject Area support coming soon.
You can find more information here
About : https://docs.oracle.com/en/cloud/paas/analytics-cloud/acubi/oracle-analytics-ai-assistant.html
Demo : https://docs.oracle.com/en/cloud/paas/analytics-cloud/acubi/generate-visualizations-oracle-analytics-ai-assistant.html
Tutorial : https://docs.oracle.com/en/cloud/paas/analytics-cloud/tutorial-oa-assistant/index.html#before_you_begin
Hi @Benjamin Arnulf-Oracle @Jacques V.-Oracle ,
Could you please provide more technical details about the AI Assistant?
I'm interested in understanding which LLM is used behind the scenes, where the model and indexes are stored, whether it will be possible to change the model in a future release, and if there are any limits on the number and size of datasets that can be indexed... This is just a start, and I’d love to learn more :)
Hi @Federico Venturin and thank you for your questions.
I will reply to the question about the possibility to change the model in a future release. The answer is yes, this is planned and added by Jacques, under safe harbor, in the official product roadmap under the line items: •Gen AI BYO-LLM integration (OpenAI) •Gen AI BYO-LLM integration (Cohere, llama 3)
I will let Jacques to reply to the other questions.
The model and the model management are entirely managed by Oracle and is not exposed to customers in any way. Our goal is to have the ability to always leverage the best model to support our use case and we plan to continuously evaluate which models perform best for the use case and adapt accordingly over time.
We optionally plan to support other models in a Bring Your Own license approach where customer provided keys will allow you to leverage public LLMs to power this use case. But the Out-of-the -box default model will not be exposed to customers.
All, please find an additional blog from Product Marketing for more information:
Will there be any details on the rollout dates? Is the rollout just OCPU based? My other question is this feature dependent on the September 2024 update?
Hi @Joel , the feature is included in the September 2024 release. Other questions might be answered by @Jacques V.-Oracle if this is allowed.
This is great! I just have a quick question—I've completed the indexing, but I don't see the Assistant option. Does this mean it's not available in our environment yet? For reference, I do have administrator access on the box.
Just curious as the link below says this >
Link : https://docs.oracle.com/en/cloud/paas/analytics-cloud/acubi/generate-visualizations-oracle-analytics-ai-assistant.html
This quarter, we’re celebrating five members of the Oracle AI & Analytics Community for their outstanding participation in the Idea Labs. Through their ideas, engagement, and collaboration, they help shape the future of Oracle AI & Analytics. By sharing thoughtful suggestions, sparking discussion, and earning community…
The Oracle AI Data Platform Keynote is now available in replay. See the following link to watch it: Youtube Video
The September 2026 Oracle AI and Analytics Product Roadmap is here, bringing the latest product direction, innovation highlights, and upcoming capabilities into focus. This quarterly update gives you a fresh look at recently delivered innovations and what is ahead across Oracle AI Data Platform, OCI Enterprise AI, Fusion…
Explore how to build agents using code in AI Data Platform. We'll create an agent that examines and analyzes the contents of a Github repository and reports its findings. The session demonstrates how to use different developer tools to create agents including the AIDP extension for VS Code and Codex. Using these tools we…
Agenda Presentation: Optimize FDI Workbook Performance Learn best practices and optimization strategies for building performant workbooks in Oracle FDI. We’ll share practical tips and lessons learned to help report authors avoid common pitfalls and design workbooks with performance in mind from the start. As a special…
Learn how to deploy an AI Data Platform agent, register it to the AI Registry, and introduce it to business users through Agent Hub. This session covers how to prepare an agent for deployment, configure endpoint access and authentication, invoke a deployed agent, and monitor it after deployment. The session also introduces…
