From Raw Agentic AI to Orchestration-First Thinking
AI agent orchestration in the enterprise is the coordinated design, governance, and execution of multiple AI agents and workflows so they can work together across systems of record, decision engines, and business processes, instead of relying on isolated, raw model power to solve complex tasks on their own. Pega’s Peter van der Putten warns that “people have maybe some magical thinking that you throw an AI model at a problem and it will sort itself out,” and cites Gartner’s prediction that over 40% of agentic AI projects will be canceled. OutSystems sees the same risk as enterprises rush into multi-agent experiments without a workflow orchestration strategy. Both vendors argue that the real competitive edge now lies in AI agent coordination, tight alignment with existing processes, and controls on cost and data access rather than in chasing the newest reasoning-heavy model.
Pega’s Customer Engagement Studio: Orchestration as the Product
Pega’s new Customer Engagement Studio makes orchestration the center of its enterprise agentic AI story. Built as a governed layer on top of Customer Decision Hub, it guides marketers from initial brief to live, personalized campaign through a single conversational interface. Under the hood, specialized agents for strategy, creative, data science, compliance, and performance are not free-roaming; they are coordinated around shared goals, guardrails, and approval flows. Van der Putten points to Wells Fargo, which runs six billion next best action decisions every month in under 250 milliseconds, to show that decisioning scale is not the bottleneck. The real challenge is organizing enough actions and content around that decisioning engine. By turning agentic AI into a structured workspace rather than a loose collection of tools, Pega aims to reduce failed experiments and shift attention from model cleverness to measurable outcomes and governed execution.
OutSystems: Neutral Orchestrator Rather Than System of Record
Where Pega builds orchestration around its own decision hub, OutSystems wants to be the neutral AI agent orchestration layer sitting above SAP, Salesforce, and other enterprise platforms. CEO Woodson Martin argues that “everybody’s aiming” at enterprise agent orchestration, but his company’s edge is not being a system of record. For years, OutSystems has acted as the “glue between commercial off-the-shelf solutions,” orchestrating processes like fund onboarding across about 80 systems for one asset management customer. That history now feeds its pitch as an agentic systems platform that focuses on workflow orchestration strategy and integration, not data ownership. CIO Tiago Azevedo stresses that the platform is agnostic to underlying systems and data stores, bringing them together “into a form that makes sense for a process.” In a crowded enterprise agentic AI market, vendor neutrality and separation from core data records become selling points.
Open Models, Shadow AI, and Cost-Governed Coordination
OutSystems extends its neutrality to models and tools. Its Agent Experience layer exposes Model Context Protocol and Agent2Agent services so developers can plug in Claude Code, Codex, Cursor, or Kiro without rewriting agent logic. Customers can swap providers or route through Amazon Bedrock while preserving AI agent coordination. Yet openness has a cost: Azevedo describes “shadow AI,” where departments spin up their own agents, driving token usage and budget anxiety. He notes that a single business value consultant in Australia consumed USD 7,500 (approx. RM34,500) in tokens every week on Anthropic models, forcing tighter control of which projects receive token budgets. In response, OutSystems emphasizes centralized orchestration of access, costs, and data scopes, while Pega moves toward outcome-based pricing instead of metering tokens. Both shifts show enterprises want control frameworks around agentic AI, not unbounded experimentation.
Enterprise Buyers Prioritize Coordination Over Agent Brilliance
The common thread between Pega and OutSystems is that AI agent orchestration, not raw model power, is becoming the main differentiator in enterprise agentic AI. Pega’s governed Customer Engagement Studio and OutSystems’ Agentic Enterprise Orchestration both show that buyers care most about consistent workflows, data access rules, and cost controls. Uncoordinated multi-agent systems risk falling into the 40% of projects Gartner expects to be canceled, either because they are too hard to govern or they never tie into real processes. By contrast, structured orchestration lets enterprises plug in new models or tools while keeping decisioning logic, compliance, and workflows stable. As agent frameworks converge, choices between platforms are less about whose agents are smarter and more about who offers neutral integration, strong governance, and a workflow orchestration strategy that scales beyond a few high-profile pilots.
