Oracle and Google Cloud announced on July 30, 2026 that they are expanding their partnership to bring Google’s Gemini models directly into Oracle’s enterprise application portfolio. The move puts agentic AI capabilities inside the business systems that tens of thousands of companies already use to run their operations.

The integration covers Oracle Fusion Cloud Applications — which includes ERP, HCM, supply chain, and CX modules — along with Oracle NetSuite. NetSuite alone serves more than 44,000 customers across 220 countries, making this one of the broader rollouts of a frontier AI model into day-to-day business software.

What’s Actually Changing

The integration works through Oracle AI Agent Studio for Fusion Applications. That is the development environment Oracle built to let companies configure, connect, and deploy AI agents on top of their ERP data. Previously, Oracle was shipping agents built on its own models. Now it is opening the door to Google’s Gemini models as an alternative.

Two Gemini models are being added. Gemini 3.1 Flash-Lite is designed for high-volume, cost-sensitive workloads — the kind of repetitive processing that runs constantly inside ERP systems. Gemini 3.5 Flash handles more complex tasks including reasoning-heavy decisions and multi-modal work such as video and presentation generation.

Chris Leone, Oracle’s EVP for applications development, framed it as a choice issue: “By bringing Gemini to Oracle AI Agent Studio for Fusion Applications, we are giving customers and partners greater choice.”

Satish Thomas, Google Cloud VP, described the goal as making Gemini accessible “in the applications and agentic workflows they rely on” — rather than requiring customers to build integrations from scratch.

Why This Is Worth Paying Attention To

The choice of delivery channel matters here. Oracle Fusion and NetSuite are not experimental tools. They handle financial close processes, payroll runs, supplier invoicing, inventory management, and customer service workflows for organizations from mid-market companies up to global enterprises.

When a frontier AI model gets embedded at that layer, it changes who has access to it. A business does not need to hire an AI team or license a separate platform. The capability arrives inside software they already have and already trust with their most sensitive financial data.

This is different from enterprise software vendors adding a chatbot layer on top of existing products. Oracle AI Agent Studio is designed for building agents that take autonomous action across workflows — agents that can read a supplier invoice, match it against a purchase order, flag discrepancies, and route exceptions for human review, without a person involved in the standard case.

Google’s models now bring expanded reasoning and multi-modal capabilities into that environment. The combination of Gemini 3.5 Flash’s reasoning and Oracle’s access to transactional ERP data in real time is the thing that could materially change what those agents can do.

The Pattern This Follows

This announcement is part of an accelerating pattern across enterprise software. Workday reported in May that more than 4,000 enterprise customers are now actively using AI agents within its HR and finance platform, with agentic AI ARR approaching $500 million. SAP has been embedding its Joule AI assistant across its product suite. Salesforce’s Agentforce platform is doing similar work in CRM.

The message from all of these announcements is consistent: enterprise software vendors are racing to become the preferred runtime for AI agents, not just the system of record for business data. They have structural advantages — existing customer relationships, established trust around sensitive data, and deeply integrated workflows — that pure-play AI vendors do not.

For Oracle, adding Gemini to its agent platform does two things. It gives customers flexibility on which AI model powers their workflows. And it signals that Oracle is competing as an AI platform, not just an ERP vendor.

What This Means for Business

For companies already on Oracle Fusion or NetSuite, this changes the cost-benefit calculation on AI automation projects. Instead of building a separate integration between a frontier model and your ERP system — which typically involves significant data plumbing work and security review — the integration is arriving pre-built inside tools you already pay for.

That removes a common barrier to getting started. The organizations most likely to move quickly are those already comfortable running agentic workflows inside Oracle AI Agent Studio, who now have access to a broader set of models for different task types.

For businesses not yet on Oracle systems, the announcement is a useful signal about where enterprise software is heading. The major ERP and CRM vendors are all betting that AI agents will become a core layer of how business operations run — and they are building the infrastructure to host those agents inside their platforms rather than letting a separate AI provider own that relationship.

A few practical implications worth tracking:

Cost efficiency at volume. Gemini 3.1 Flash-Lite is explicitly positioned for high-efficiency, high-volume workloads. For companies processing large numbers of invoices, contracts, or records daily, the model choice will directly affect the cost of running those agents at scale.

Multi-modal workflows become accessible. Gemini 3.5 Flash’s ability to reason across video and presentations opens use cases that were not realistic inside ERP systems before — supplier product demonstrations, compliance training documentation, visual inspection workflows in supply chain.

The AI skills gap does not go away. Embedding better models inside Oracle does not automatically mean organizations will use them well. Designing effective agents, defining the right decision points for human oversight, and interpreting agent outputs all require people who understand both the business process and the AI system. That combination is still uncommon in most organizations.

The Bigger Picture

The companies gaining the most from enterprise AI right now are not the ones who found the best standalone AI tool. They are the ones who figured out how to wire AI into their operational workflows — the daily processes where decisions happen, records get updated, and work actually gets done.

Oracle embedding Gemini into Fusion and NetSuite is another step toward that integration becoming the default rather than a project. For business leaders, the question is no longer whether the technology will be available in their business software. It increasingly is. The question is whether their teams have the skills and the process clarity to use it.

---

Understanding how AI agents operate inside enterprise systems — and what your team needs to work with them effectively — is core to what Enterprise DNA teaches. Explore our AI and data training programs or talk to us about deploying AI workflows in your business.