Microsoft recently announced an evolution of its AI stack: Wave 3 of Microsoft 365 Copilot, model-agnostic support for Claude, and the upcoming May 1 launch of Agent 365 and the M365 E7 Frontier Suite. These moves are part of a broader "Frontier Transformation," the next step in the evolution of the Frontier Firm. This new focus centers on two essential pillars: Intelligence, which ensures AI remains context-aware and deeply integrated into your workflows, and Trust, which delivers the enterprise-grade security, governance, and guardrails required for global scale. The progression represents a holistic push to integrate advanced AI with enterprise-grade security across the Microsoft 365 ecosystem, offering IT departments the robust controls required to manage a new era of "agentic" software.
Key Points of the Frontier Transformation
Wave 3 of Copilot: Evolves Copilot into an autonomous agent powered by Work IQ, which uses organizational data to take initiative and provide context-aware results.
Copilot Cowork: A multi-step "autopilot" that coordinates complex tasks across Word, Excel, and PowerPoint, such as drafting a proposal from a sales report.
Model Diversity (Claude Integration): Through the Frontier preview program, Copilot is now model-agnostic and integrates Anthropic’s Claude alongside OpenAI GPT models to automatically select the best engine for specific tasks.
Agent 365: A centralized control plane and registry in the Microsoft Admin Center used to govern, monitor, and secure all enterprise AI agents via Microsoft Entra.
Microsoft 365 E7: A new "Frontier Suite" ($99 per user/month) launching May 1, 2026, that bundles E5, Copilot, and Agent 365 into a single, secure SKU.
The Intelligence Pillar: Microsoft 365 Copilot Wave 3
The "Wave 3" update transforms Microsoft 365 Copilot from a passive assistant into an autonomous "agent" capable of taking initiative within daily applications.
Work IQ and Contextual Awareness
At the heart of this intelligence is Work IQ, Microsoft’s organizational intelligence layer. Work IQ aggregates enterprise knowledge, including documents, communications, and collaboration patterns, to inform AI outputs. It allows Copilot to understand company-specific jargon, reference recent project details, and adhere to corporate style guidelines. For IT, this means AI tools automatically recognize context, such as team structures and shared files, without manual data integration.
Copilot Cowork and Multi-Step Autopilot
Wave 3 introduces Copilot Cowork, a feature that allows the AI to execute complex, multi-step actions over time. Users can issue high-level commands, such as drafting a client proposal by gathering data from a sales report and generating a summary presentation. Copilot coordinates these actions across Word, Excel, and PowerPoint, keeping the user updated and allowing for manual intervention or corrections. Critically, Cowork operates entirely within the organization’s governed environment, storing all created files in OneDrive or SharePoint to ensure compliance.
Expanded Model Diversity
Microsoft has shifted Copilot to a model-agnostic backend, meaning it no longer relies on a single LLM. Through the Frontier preview program, Microsoft 365 Copilot now integrates Anthropic’s Claude alongside the latest OpenAI GPT models. The system automatically selects the optimal model for a specific task behind the scenes. This approach prevents vendor lock-in and allows IT leaders to benefit from industry-wide AI innovations without managing multiple platforms.
The Trust Pillar: Microsoft Agent 365
To combat "agent sprawl," where unmanaged AI bots and scripts create security blind spots, Microsoft introduced Agent 365. This brand-new product serves as a unified control plane for AI agents across the enterprise.
Centralized Governance and Identity
Agent 365 provides a central registry that enables IT to monitor, manage, and secure every AI agent running in the organization. Technically, agents are registered as first-class entities in Microsoft Entra (formerly Azure AD), giving them unique identities similar to service accounts.
Security Integration
Agent 365 plugs directly into Microsoft Defender, Entra, and Purview. This ensures that all agent activities, such as accessing a sensitive file or sending an email, fall under the same security blanket and audit rules as human users. If an agent behaves unexpectedly, IT can audit its history or disable it instantly, maintaining total oversight of the automation landscape.
Licensing the Future: Microsoft 365 E7
To simplify the procurement of these advanced tools, Microsoft introduced a new top-tier plan: Microsoft 365 E7, also known as "The Frontier Suite."
The All-in-One AI Bundle
Available starting May 1, 2026, the E7 suite combines the capabilities of the current E5 plan with the new AI tools:
- Microsoft 365 E5: Core productivity, Windows, and advanced security.
- Microsoft 365 Copilot: The full agentic AI experience across Office apps.
- Agent 365: The governance and management portal for AI agents.
- Advanced Security: Includes the full Entra identity suite, Intune, and Purview.
The E7 suite is priced at $99 per user/month. This pricing is designed to be more cost-effective than purchasing E5, Copilot, and Agent 365 separately. For IT departments, E7 eliminates the "procurement puzzle" by providing a single SKU that ensures all components are technically aligned and secure by default. Find out more about M365 E7.
Strategic Implications for IT Professionals
The Frontier Transformation represents a shift in how IT departments operate. With projections suggesting 1.3 billion AI agents will be in circulation by 2028, managing AI is becoming as common as managing mobile devices.
From Pilots to Production
Microsoft emphasizes that organizations are ready to move past "AI hype" and experimentation. Frontier Transformation provides the infrastructure to operationalize AI company-wide, drawing on Microsoft's internal experience supporting over 500,000 agents.
Cultural and Ethical Responsibility
IT leaders now play a central role in the ethical use of AI. This includes: updating training to help employees effectively use "AI co-workers," establishing policies on who can build agents using Agent Builder or Copilot Studio, working with HR and legal teams to review AI outputs for bias or errors, and ensuring human oversight remains in place for critical decisions.
Operational Readiness
To prepare for the General Availability of these products on May 1, 2026, IT teams should: evaluate the cost-benefit of consolidating licenses into the E7 suite, skill up on AI administration, including configuring data sources for Copilot and using the Agent 365 registry, and join the Frontier program for early access to agentic features like Copilot Cowork.
Frontier Transformation is not just a product launch, but the beginning of a journey into AI-powered work where intelligence and trust operate in tandem.
