Enterprise AI
ERP software just started
letting you pick your AI
Oracle has embedded Gemini directly inside its business applications. The same company only added OpenAI models in June — adding a second frontier model just two months later says less about Oracle's product roadmap and more about what ERP vendors are becoming.
Gemini 3.5 Flash just moved
inside business applications
On July 30, 2026, Oracle officially announced an expanded partnership with Google Cloud to bring Gemini models to its enterprise applications customers. The rollout targets Oracle Fusion Applications and Oracle NetSuite, offering two models: cost-efficient Gemini 3.1 Flash Lite, and Gemini 3.5 Flash for more complex reasoning and video/presentation generation.
The integration point is the newly launched Oracle AI Agent Studio for Fusion Applications — a development platform for building AI automation by combining Oracle-native, partner, and external agents. The move was also covered in Japan by Innovatopia.
The models and their intended use
| Model | Intended use |
|---|---|
| Gemini 3.1 Flash Lite | High-frequency, low-cost routine tasks |
| Gemini 3.5 Flash | Complex reasoning, video & presentation generation |
Why this is more than
a feature update
The role of the ERP vendor itself is shifting, not just its feature list.
Enterprise software has traditionally shipped with one AI vendor baked in, chosen and fixed by the vendor. But Oracle layered Gemini onto the very same Cloud footprint where it started offering OpenAI models just 49 days earlier. This doesn't read as Oracle picking a "best" model — it reads as a deliberate choice to let each customer decide for themselves. Oracle is shifting from being the implementer of a single AI to becoming a switchboard for multiple models, and this partnership is one of the first concrete instances of that shift.
The ERP vendor's choice is no longer which AI.
It's building the switchboard that lets you choose.
Who benefits, and how
PMs, sales, and ops staff
You can now call Gemini from the same Oracle screens you already use, cutting out one more app switch. Approvals and transaction execution still run through Oracle's own workflows.
Developers and SI engineers
AI Agent Studio lets you combine Oracle-native and external agents into a single automation. A multi-model architecture is now the default assumption, not an afterthought.
Marketers and planners
Gemini 3.5 Flash's video and presentation generation becoming available inside the business app itself could shorten the deliverable-creation loop.
What to check next
If your organization already runs Oracle Fusion or NetSuite, it's worth doing more than just welcoming the convenience:
Confirm your rollout timeline
Ask your Oracle account team when Gemini 3.1 Flash Lite / 3.5 Flash actually activates under your contract.
Match models to use cases
Use Flash Lite for routine tasks and Flash for content generation or complex reasoning, balancing cost against accuracy.
Watch competing ERP vendors
Whether SAP or Workday pursue a similar multi-model strategy will indicate whether this becomes an industry standard.
The lock-in hiding behind "choice"
Being able to pick between models looks, at first glance, like reduced vendor lock-in. But what you're actually locked into isn't the model — it's Oracle's orchestration layer, AI Agent Studio. Swap the model, and the workflow, approval chain, and data plumbing above it remain thoroughly Oracle-specific.
Running both OpenAI's and Google's models inside a single ERP also effectively doubles the surface area for data governance and compliance audits. Whether an organization can precisely track which data reaches which model depends entirely on how carefully it designs its own operating rules. Separate billing lines for each model provider also make consolidated visibility into monthly AI spend a real practical question worth resolving before rollout.