Daily Enterprise Agentic AI Report - 2026-05-18

• The most direct competitive news for the radar is Freshworks Freddy AI Agent Studio in Freshservice: an explicit "AI Agent Studio" for service management, with no-code, prebuilt agents, MCP Gateway, governance, and measurement.
• Coupa Compose confirms that market language is moving toward "Agentic-as-a-Service": agent packages, outcome-based pricing, and deployment services for procurement, finance, and supply chain.
• UiPath reinforces an important enterprise thesis: coding agents are not enough unless they connect to CI/CD, testing, security, deployment, observability, and governance.
• Symphony and SailPoint show two critical verticals: agents for regulated financial markets and identity control for agents/non-human identities.
• In capabilities, Microsoft MDASH and the joint Five Eyes guidance reinforce that architecture, identity, permissions, traceability, and agent evaluation are now first-order criteria, not secondary details.

1. Freshworks presents Freddy AI Agent Studio in Freshservice

• URL: https://www.freshworks.com/pressrelease/freshworks-unveils-ai-agent-studio-in-freshservice-to-unlock-service-transformation-that-drives-compounding-business-growth/
• Publication date: 2026-05-14
• Summary: Freshworks announces a no-code AI Agent Studio inside Freshservice to create custom agents, use prebuilt domain agents, and extend them with agentic workflows. It includes integration with Microsoft Teams, Slack, employee portals, HRIS tools such as Workday and Rippling, and MCP Gateway to access external context from tools such as Notion, ClickUp, and Linear.
• Relevance versus Oracle AI Agent Studio: Very high. It is a direct competitor in the "enterprise agent studio" category, although focused on ServiceOps, ITSM, and employee service rather than the full Fusion ERP/HCM/SCM/CX footprint. Potential impact: Strong signal. Freshworks is positioning agility, fast implementation, and lower complexity against legacy platforms, a narrative that can compete with Oracle in mid-market accounts and service areas.
• Detected capabilities: No-code studio, domain agents, agentic workflow library, MCP Gateway, embedded governance, AI Insights, xLAs, integration with enterprise tools, and experience measurement.

2. Coupa launches Coupa Compose and Catalyst for Agentic-as-a-Service

• URL: https://www.prnewswire.com/news-releases/coupa-launches-coupa-compose-and-catalyst-to-accelerate-agentic-ai-value-and-delivery-at-inspire-2026-302769893.html
• Publication date: 2026-05-12
• Summary: Coupa announces Coupa Compose as an environment to build, manage, and orchestrate a digital workforce of agents across procurement, finance, and supply chain. The package combines agents, outcome-oriented pricing, and transformation services with forward-deployed engineers and architects.
• Relevance versus Oracle AI Agent Studio: High. Coupa competes in business processes close to procurement, spend, payments, supply chain, and finance, areas where Oracle Fusion also has a strong proposal. Potential impact: Strong signal. The "Agentic-as-a-Service" approach may become a purchasing category: not only a platform, but an operating package to capture value quickly.
• Detected capabilities: Agentic orchestration, governance over data/workflows/agents, digital workforce, outcome-based pricing, adoption services, and proprietary spend data as grounding.

3. UiPath integrates coding agents into its enterprise orchestration platform

• URL: https://www.uipath.com/newsroom/uipath-for-coding-agents-launch
• Publication date: 2026-05-12
• Summary: UiPath announces native integration for coding agents, starting with Claude Code and OpenAI Codex, with the goal of turning agent outputs into governed, tested, deployed, and operated enterprise automations at scale.
• Relevance versus Oracle AI Agent Studio: High. UiPath does not compete as an ERP/HCM suite, but it does compete as an orchestration, automation, and process governance layer over enterprise systems. Potential impact: Strong signal. It reinforces that the competitive advantage is not only in the model, but in turning agents into safe operational execution.
• Detected capabilities: Orchestration, testing, deployment, observability, governance, multi-agent compatibility, CI/CD, security policies, and support for both technical and business users.

4. Symphony presents AI Agent Studio for financial markets

• URL: https://symphony.com/company/newsroom/press-release/symphony-introduces-its-ai-agent-studio-at-innovate-2026/
• Publication date: 2026-05-12
• Summary: Symphony shows its AI Agent Studio at Innovate 2026, focused on creating agents with Symphony tools, integrating customer agents into Symphony workflows, and using a library of first-party and partner MCPs, with the option to bring your own MCPs and LLMs.
• Relevance versus Oracle AI Agent Studio: Medium-high. It is a vertical alternative for financial markets and regulated communications, not a horizontal suite, but it competes on the same criteria of security, context, integration, and governance. Potential impact: Moderate-strong signal. It indicates that vertical agent studios are specializing by regulated industry.
• Detected capabilities: AI Agent Studio, external agent integration, MCP library, BYO MCP/LLM, AI insights, and a security framework for financial markets.

5. SailPoint launches Agentic Fabric to secure agent identities

• URL: https://www.sailpoint.com/press-releases/sailpoint-launches-agentic-fabric
• Publication date: 2026-05-11
• Summary: SailPoint presents Agentic Fabric to discover, map, govern, and protect agent identities, machine identities, and applications. The proposal links each agent with a human owner, data, systems, access policies, and least-privilege controls.
• Relevance versus Oracle AI Agent Studio: Medium-high. It is not an agent builder, but it provides a critical layer for any enterprise agent deployment, including comparison with Oracle on security, accountability, and controls. Potential impact: Strong signal. Agent identity is becoming an essential category for regulated enterprise adoption.
• Detected capabilities: Agent discovery, identity relationship map, human ownership, lifecycle controls, access policies, zero-standing privilege, just-in-time access, and shadow AI discovery.

1. OpenAI brings Codex to mobile and strengthens remote agentic work

• URL: https://openai.com/index/work-with-codex-from-anywhere/
• Publication date: 2026-05-14
• Summary: OpenAI announces Codex in the ChatGPT mobile app, with the ability to review outputs, approve actions, redirect work, start tasks, and move between connected hosts while Codex keeps working on Mac or remote environments. The announcement also mentions access tokens for Enterprise/Business and HIPAA support in eligible local environments.
• Relevance for enterprise agentic solutions: High. It introduces a human-agent collaboration pattern for long-running work: the agent operates in the background and the human intervenes from anywhere at decision points. Potential impact: Strong signal. For enterprise platforms, mobility and remote approvals can accelerate adoption, but they increase the need for risk policies, approvals, and auditability.

2. Anthropic and PwC expand partnership to deploy Claude across enterprise functions

• URL: https://www.anthropic.com/news/pwc-expanded-partnership
• Publication date: 2026-05-14
• Summary: Anthropic and PwC announce an expanded alliance to use Claude in technology development, deal execution, and reinvention of enterprise functions. PwC will deploy Claude Code and Cowork, create a Center of Excellence, and train/certify 30,000 professionals.
• Relevance for enterprise agentic solutions: High. Global integrators are a key channel for taking agents from pilots to process transformation and scaled deployment. Potential impact: Strong signal. Oracle, SAP, Salesforce, and ServiceNow will compete not only against products, but against implementation ecosystems that package agents with consulting and operating change.

3. Microsoft MDASH demonstrates orchestration of more than 100 agents for code security

• URL: https://www.microsoft.com/en-us/security/blog/2026/05/12/defense-at-ai-speed-microsofts-new-multi-model-agentic-security-system-tops-leading-industry-benchmark/
• Publication date: 2026-05-12
• Summary: Microsoft presents MDASH, a multi-model agentic system for vulnerability discovery and remediation. It orchestrates more than 100 specialized agents and found 16 new vulnerabilities in Windows components, with notable results on CyberGym and internal testing.
• Relevance for enterprise agentic solutions: Very high. MDASH shows that important advances may come from specialized multi-agent systems, not only from a more powerful model. Potential impact: Strong signal. It reinforces the need for architectures with specialized agents, debate/validation, deduplication, exploitability testing, and traceability.

4. NSA/Five Eyes publish joint guidance for careful adoption of agentic services

• URL: https://www.nsa.gov/Press-Room/Cybersecurity-Advisories-Guidance/
• Publication date: 2026-04-30; analysis and follow-up active in May 2026
• Summary: The NSA advisories page lists the joint guidance "Careful Adoption of Agentic AI Services", issued by Five Eyes cybersecurity agencies. The document has become a reference for assessing the risks of autonomous agents with permissions over real systems.
• Relevance for enterprise agentic solutions: Very high. Any platform like Oracle AI Agent Studio will need to demonstrate identity, least privilege, human approvals, traceability, adversarial testing, and recovery. Potential impact: Strong signal. Agent security is moving from technical recommendation to regulatory and compliance expectation for critical sectors.

5. Microsoft and OpenAI reinforce the governed agentic development pattern

• URL: https://help.openai.com/en/articles/10128477-chatgpt-enterprise-edu-release-notes
• Publication date: 2026-05-14
• Summary: The Enterprise/Edu release notes detail Codex remote access from mobile, access tokens for non-interactive local workflows, and governance surfaces for token activity. The experience shows live context from the environment where Codex operates, including approvals, screenshots, terminal, diffs, and test results.
• Relevance for enterprise agentic solutions: High. It points toward the control requirements ahead: agents connected to real environments, but auditable and governed by workspace. Potential impact: Moderate-strong signal. This pattern can extend to non-technical agents: operations, finance, sales, HCM, and supply chain will also need approvals, logs, and centralized policies.

1. Oracle publishes Learning Path for Fusion AI Agent Studio

• URL: https://blogs.oracle.com/fusioncoe/fusion-ai-agent-studio-learning-path
• Publication date: 2026-04-30; crawled in May 2026
• Summary: Oracle publishes a learning path for Fusion AI Agent Studio that organizes concepts, security, hierarchical agents, workflow agents, tools, monitoring, evaluation, traceability, MCP, A2A, APIs, OAuth2, and prompt engineering. Competitive signal: Strong. Against SAP, Freshworks, or ServiceNow, Oracle needs to demonstrate not only product, but clear playbooks for building, operating, and optimizing agents.
• Impact for Oracle Fusion customers: High. It reduces adoption friction by turning AI Agent Studio into a learnable discipline with technical resources and implementation guides.

2. Oracle explains MCP and A2A in Fusion Applications Release 26A

• URL: https://blogs.oracle.com/fusioninsider/what-you-need-to-know-about-mcp-a2a-in-fusion-apps
• Publication date: 2026-03-11; crawled in May 2026
• Summary: Oracle explains how Release 26A incorporates MCP for structured access to external data/tools and A2A for secure collaboration between Fusion agents and external agents. Both capabilities are implemented in AI Agent Studio with Fusion Applications security and role-based access controls. Competitive signal: Strong. Interoperability with MCP/A2A is key to competing with Freshworks MCP Gateway, Symphony MCP library, SAP agent-to-agent, and Microsoft/ServiceNow control layers.
• Impact for Oracle Fusion customers: High. It enables a shift from isolated agents to multi-agent and multi-system workflows without abandoning Fusion security controls.

3. Oracle Readiness 26B confirms agentic app creation from AI Agent Studio

• URL: https://docs.oracle.com/en/cloud/saas/readiness/common/26b/common26b/26B-common-wn-f43462.htm
• Publication/crawl date: Release 26B, crawled in May 2026
• Summary: The documentation indicates that agentic apps can now be created in AI Agent Studio with specialized agent teams, unified experiences, information displays, Ask Oracle advisors, actions, and communications. Competitive signal: Strong. It is a direct response to the builder and studio market: Oracle is documenting agentic app creation as a product capability.
• Impact for Oracle Fusion customers: Very high. It confirms the move from individual agents to agentic applications composed of specialized agents.

4. Oracle introduces Fusion Agentic Applications for Customer Experience

• URL: https://www.oracle.com/news/announcement/oracle-introduces-fusion-agentic-applications-for-cx-2026-04-09/
• Publication date: 2026-04-09
• Summary: Oracle announces five Fusion Agentic Applications for CX: Contract Compliance Workspace, Cross-Sell Program Workspace, Marketing Command Center, Sales Command Center, and Service Manager Workspace. They operate within the Fusion security framework and rely on Agentic Applications Builder. Competitive signal: Strong. Oracle competes with Salesforce Agentforce, ServiceNow CRM/service agents, and Freshworks CX/ServiceOps agents in revenue and service domains.
• Impact for Oracle Fusion customers: High. It grounds AI Agent Studio in commercial and service use cases with continuous monitoring, risk analysis, recommended actions, and manual work reduction.

5. Oracle Workflow Agents combine deterministic control and autonomous intelligence

• URL: https://blogs.oracle.com/fusioncoe/introducing-workflow-agents
• Publication date: 2026-02-12; crawled in May 2026
• Summary: Oracle describes Workflow Agents as AI Agent Studio agents for executing end-to-end workflows by combining deterministic control, reasoning, memory, coordination, human-in-the-loop, and multi-agent collaboration. Competitive signal: Moderate-strong. It reinforces the "built in, not bolted on" message and addresses the main enterprise concern: autonomous agents without control or auditability.
• Impact for Oracle Fusion customers: High. This capability is especially relevant for auditable processes with exceptions: unexpected documents, incomplete data, approvals, recalls, payroll, and regulated operations.