Oracle’s Applications Analyst Summit in April 2026 put a decade of “multiple ERP paths” messaging under pressure. Oracle leadership drew a hard boundary around AI: The agentic future is built for Fusion SaaS, not for E-Business Suite (EBS), JD Edwards, PeopleSoft, or Hyperion. Rondy Ng, executive vice president of application development, said it plainly: “These agents are not coming to [EBS]. They’re not coming to our on-premises solutions, to JD Edwards, to PeopleSoft, to Hyperion. To enable this, you have to get to SaaS.” AI is the dividing line. Fusion is where Oracle is building the next operating model.

What Changed At The Summit

  • Oracle pivoted from “agents” to productized agentic apps. Oracle is not pitching headless assistants. It is shipping named agentic apps that bundle specialized agents with standard operating procedures, thresholds, and audit trails. A ledger agent is in production. Consolidation and financial planning/analysis agentic apps were demoed live. Collections workspace and channel revenue agentic apps are expected this year. This is not a feature add; it is a workflow and controls model for finance.
  • Oracle ended the multiproduct hedge for AI. The prebuilt agents, Agent Studio, and the emerging AI Agent Marketplace are positioned around Fusion. Premier Support may run through 2036 for legacy products, but the AI-led innovation story is SaaS-first in both messaging and product packaging.
  • Oracle made migration AI-first, not AI-after. Oracle introduced an enterprise configuration agent that ingests EBS, PeopleSoft, and JD Edwards metadata and proposes a Fusion enterprise structure with options and trade-offs. Work that took systems integrators months now runs in hours. Oracle also restructured customer success programs to activate agents from day one of Fusion, not after stable operations. Customers no longer migrate first and adopt AI later. That accelerates value, but it increases implementation risk if controls and process ownership are not ready.
  • Agent supply is strong. Adoption proof is not. Oracle cited 600 prebuilt agents, 100-plus partner agents, and 32,000 certified developers. What was missing: production deployment counts, activation rates, and business outcome metrics. That is the key gap in Oracle’s AI narrative today.
  • Pricing messaging held. Consumption reality did not. Oracle reiterated no plan to charge for embedded agents in Fusion. The unresolved question is renewal-time pricing once agent-heavy workloads scale. Oracle did not commit to a model on stage.
  • Customer references skewed regulated, capital-intensive industries. AT&T and Northwell Health led fireside chats. Guardian Life and Equinix appeared in ERP interviews. These industries benefit when Oracle can span data, infrastructure, and applications. Notably, Oracle did not address healthcare-scale progress questions that still dominate many CIO conversations.

What This Means For CIOs

CIOs running Oracle’s on-prem ERP should treat Fusion as the only AI innovation path and negotiate a commercial and technical bridge to Fusion AI in the next 12–18 months. Waiting for a full migration will compound security, talent, and productivity gaps.

What To Do Now

  • Use 2036 as leverage, not runway. Ask: “What is the commercial bridge that lets us consume Fusion AI before completing the SaaS migration?” The answer, or the dodge, sets your negotiating position.
  • Demand renewal guardrails. Even if today is “no charge,” lock in definitions, caps, audit rights, and protections against surprise consumption pricing.
  • Pilot with controls first. If you test agentic apps, require audit trails, thresholds, and human override paths as entry criteria.
  • Sequence for risk. AI-first activation can accelerate value, but it can also compound implementation risk. Treat agent rollout as an operating model change, not a UI upgrade.

Want to pressure-test Oracle’s SaaS-only AI stance and define the guardrails you’ll need to adopt Fusion agents safely? Clients can schedule a guidance session with us.