Power Platform 2026 Wave 1: Your Complete Guide to the Latest Updates
Microsoft has released its 2026 Release Wave 1 plan for Power Platform, covering hundreds of new features rolling out between April and September 2026. If your organisation builds apps, automates processes, runs business portals, or uses AI agents on the Microsoft stack, there’s something in this wave that affects how you work.
We’ve gone through all six product areas and pulled out what matters most. Here’s what’s coming.
Key Takeaways
- The central theme across the entire wave is agentic AI. Agents that build, monitor, fix, and run processes autonomously, not just assist humans doing them manually
- Power Apps gets a modernised default UI, better offline capability, and smarter search, rolling out automatically for most users
- Power Pages adds a Security Agent that actively monitors and protects your business portals from threats including phishing and DDoS attacks
- Power Automate introduces self-healing desktop flows that automatically adapt when underlying systems change. A significant step forward for RPA reliability
- Copilot Studio deepens its integration with Microsoft 365 Copilot, Microsoft Foundry, and Work IQ, making it easier to build and scale enterprise-grade agents
- Dataverse connects business data directly into M365 Copilot experiences through Work IQ, enabling natural language access to line-of-business data
- Governance gets a major upgrade. AI-powered monitoring agents, real-time risk assessment, granular Copilot credit tracking, and automated licence reclaim
- Production deployment begins April 1, 2026. Some features turn on automatically, others require admin configuration
Power Apps: A Modernised Experience
Power Apps has been building toward a modernised interface for some time. In 2026 Wave 1, that modern look becomes the default for everyone.
A Cleaner, More Consistent UI
Model-driven apps get a refreshed header and navigation that reaches general availability this wave. For organisations running multiple model-driven apps, this means a consistent, streamlined experience across the board without any rebuild work required.
Better Offline and Mobile Capability
Canvas apps gain real-time Dataverse access in online mode alongside improved offline resilience. Makers can now configure advanced offline profiles using FetchXML and enable push notifications in custom-branded apps, meaning field-based workers get a more reliable mobile experience even in low-connectivity environments.
Smarter Search
Search in model-driven app grids and lookups gets faster and more intelligent this wave. For users working with large datasets, this reduces the time spent hunting for the right record.
AI-Powered Generative Pages
Generative pages, which use AI to automatically build app pages based on your data, expand their availability and localisation in this wave. Teams can build and scale intelligent apps faster without hand-coding every screen.
Power Pages: Smarter Portals, Stronger Security
Power Pages is the tool organisations use to build external-facing business portals for customers, partners, employees, or citizens. This wave focuses on two things: better developer tooling and significantly stronger security.
A Security Agent That Monitors Your Site
The headline addition is the Security Agent, reaching general availability this wave in two forms.
The Security Agent for site creators helps during the build process, guiding makers through configuring roles and permissions correctly, catching misconfigurations before they become vulnerabilities.
The Security Agent for admins operates continuously in production, monitoring live portals for threats including phishing attempts, DDoS attacks, spam, and offensive content. When it detects an issue, it flags it and suggests remediation.
What this looks like in practice: An insurance company runs a customer self-service portal on Power Pages. Rather than relying on periodic manual security reviews, the Security Agent monitors the site continuously. When it detects an unusual spike in traffic consistent with a DDoS pattern, it alerts the admin team immediately, well before it becomes a service disruption.
End-User Transaction Auditing
Organisations can now capture and track end-user activity on Dataverse records through Power Pages, providing a complete audit trail of what customers or partners did on the portal and when. For regulated industries, this is a meaningful compliance improvement.
Better Authentication Controls
Portal authentication gets enhanced identity management this wave, including support for encrypted ID tokens. Admins also gain granular control over which external authentication providers are permitted for their portals, useful for organisations with specific identity governance requirements.
Site Analytics and Server Log Exports
Site usage, performance data, and operational events can now be exported to your preferred analytics system. Organisations get the visibility needed to monitor portal health, diagnose issues, and report on usage without being limited to what’s available natively in the Power Pages interface.
Power Automate: Self-Healing Flows and Smarter Automation
Power Automate covers cloud flows, desktop flows (RPA), and process mining. This wave brings meaningful advances to all three with the most significant changes coming to desktop flows.
Self-Healing Desktop Flows
This is the standout update for anyone running robotic process automation. Desktop flows can now automatically detect when an underlying application or system has changed and adapt. Rather than a flow failing silently when a UI element moves or a field is renamed, the system identifies the issue and attempts to resolve it.
What this looks like in practice: A finance team runs a desktop flow that extracts data from a legacy system every morning. After a system update changes the layout of a screen, the flow would previously have failed and required a developer to manually remap the affected steps. With self-healing capability, the flow detects the change, adapts, and continues running with a notification to the admin rather than a broken process.
AI Agent Authoring for Desktop Flows
Makers building desktop flows now get AI agent assistance during the authoring process helping optimise flow design, suggest improvements, and identify potential failure points before deployment. This lowers the barrier for less experienced makers and speeds up development for seasoned ones.
Copilot Studio-Powered Actions in Cloud Flows
Cloud flows can now call Copilot Studio agents directly as actions within a workflow. This means AI agent capabilities (reasoning, knowledge retrieval, dynamic responses) can be embedded into automated processes without custom development.
What this looks like in practice: A customer onboarding flow automatically triggers when a new application is received. At a specific step, it calls a Copilot Studio agent to assess the application against policy criteria and generate a recommendation. The recommendation is logged to Dataverse and the flow continues with AI embedded in the process rather than bolted on alongside it.
Real-Time Collaboration and Version Control for Desktop Flows
Desktop flow makers gain real-time collaboration capability this wave. Multiple people can work on the same flow simultaneously. Combined with enhanced version control and centralised credential management, this makes desktop flow development significantly more manageable at scale.
Object-Centric Process Mining — Now Generally Available
Process mining reaches general availability this wave with a new object-centric model that handles complex, interconnected processes more accurately than traditional event-log approaches. A redesigned process intelligence studio workspace adds custom KPIs, flexible layouts, and native Microsoft Fabric integration giving operations and analytics teams much richer insight into how work actually flows through the organisation.
Microsoft Copilot Studio: Deeper Integration, More Powerful Agents
Copilot Studio is Microsoft’s platform for building AI agents and agentic workflows. This wave focuses on making agents easier to extend, more powerful to run, and better governed at scale.
Extending Agents Built in Microsoft 365 Copilot
Organisations using Agent Builder in Microsoft 365 Copilot can now further customise those agents through Copilot Studio adding new knowledge types, more sophisticated tools, and evaluation capabilities. This is a meaningful bridge between the M365 Copilot world and the full Copilot Studio platform.
What this looks like in practice: A financial services organisation builds a basic agent through Agent Builder in M365 Copilot to help staff find policy information. As requirements grow, they extend that agent in Copilot Studio connecting it to a broader knowledge base, adding tools that can query Dataverse records, and setting up evaluations to test response quality before deployment. One agent, progressively enhanced.
High-Value Out-of-the-Box AI Actions
New pre-built AI actions in workflows make it easier to apply AI to common automation scenarios without custom development. Teams can access intelligent capabilities (document processing, summarisation, classification) directly within their agent workflows.
Multi-Agent Orchestration
Multiple agents can now be orchestrated together, each responsible for a specific part of a workflow, handing off to the next when their task is complete. This enables more complex, end-to-end automated processes that would previously have required significant custom development.
Deeper Integration with Microsoft Foundry and Work IQ
Copilot Studio’s integration with Microsoft Foundry and Work IQ means agents can now access the latest AI models alongside your organisation’s operational data making responses and decisions more accurate, more contextual, and more specific to your business.
Governance at Scale
Enterprise governance capabilities for Copilot Studio expand significantly with more visibility into agent activity, real-time risk assessment, and controls that allow IT to set guardrails without blocking maker productivity.
Microsoft Dataverse: Business Data in the Flow of Work
Dataverse is the data platform underpinning Power Platform. This wave focuses on making that data more accessible to AI agents and easier to manage at scale.
Work IQ: Business Data Inside M365 Copilot
The headline Dataverse update this wave is Work IQ which integrates business data from Dataverse directly into Microsoft 365 Copilot experiences. Teams can access line-of-business data and trigger multi-step workflows through natural conversation in M365 Copilot, without switching to a separate application.
Autonomous agents built on Dataverse make decisions grounded in your organisation’s specific data, with adaptive learning that improves over time and full auditability through transparent agent identities.
What this looks like in practice: A relationship manager asks M365 Copilot about a customer’s recent activity. Rather than switching to Dynamics 365, the answer comes back immediately, drawn from live Dataverse records, along with a suggested next action. The whole interaction happens within the Teams or Outlook interface they’re already working in.
Developer Extensibility: APIs, MCP Servers, and Python SDK
Developers get significantly more ways to build with Dataverse this wave. Work IQ APIs and MCP (Model Context Protocol) servers allow developers to configure, extend, and integrate business agents into custom applications and workflows. A new Python SDK provides a native development experience for rapid prototyping, lowering the barrier for data engineers and developers who prefer Python over traditional .NET approaches.
Storage Management: Data Hub and Storage Advisor
Admins get new tools for understanding and managing Dataverse storage consumption. Data Hub and Storage Advisor provide insights and recommendations at the tenant level helping organisations avoid unexpected storage overruns and optimise how data is retained and organised.
Governance and Administration: Control Without Slowing Things Down
For IT teams managing Power Platform at scale, this wave delivers the most comprehensive governance upgrade in recent memory.
AI-Powered Governance Agents
Governance itself becomes agentic in this wave. AI agents now continuously monitor your Power Platform tenant. Rather than administrators running periodic manual reviews, the platform watches itself and alerts when action is needed.
What this looks like in practice: An IT admin responsible for a large Power Platform estate gets an alert that three flows in a non-production environment are connecting to an external service that recently had a known security issue. The governance agent identified the dependency, flagged the risk, and suggested remediation steps before anyone was affected.
Real-Time Risk Assessment in Copilot Studio
Before an agent is deployed, real-time risk assessment catches security and compliance issues. This means problems are identified at the point of creation rather than discovered after go-live, significantly reducing the cost and disruption of remediation.
Granular Copilot Credit Tracking and Consumption Caps
Organisations can now track Copilot credit consumption at a granular level and set configurable caps preventing unexpected budget overruns. For organisations managing Copilot rollouts across large user bases, this provides the cost predictability that finance teams need before committing to broader deployment.
Automated Licence Reclaim
Unused licences are automatically identified and returned to the pool. For large organisations where licence sprawl is a real cost issue, this feature pays for itself quickly recovering spend on licences that are assigned but not being used.
Connector Dependency Visibility
Admins can now see exactly which flows and apps depend on which connectors, making it significantly easier to assess the blast radius when a connector has a security issue or breaking change. Previously this required manual investigation across potentially hundreds of flows.
GitHub Integration and Deploy from Git
Application lifecycle management matures in this wave with GitHub integration and deploy-from-Git capability. Changes to Power Platform solutions can be tracked, reviewed, and deployed with full audit trails bringing enterprise software development practices to the low-code world.
If you want to find out how to get the most out of these updates, get in touch!
FAQs
When does Power Platform 2026 Release Wave 1 go live?
Production deployment begins April 1, 2026, with features rolling out regionally through to September 2026.
Will the new Power Apps UI change automatically for our users?
Yes. The modernised look for model-driven apps is being enabled automatically as part of this wave. We recommend reviewing the specific changes ahead of April so your teams know what to expect.
What is Work IQ and do we need a separate licence for it?
Work IQ is the integration layer that connects Dataverse business data into Microsoft 365 Copilot experiences. Licensing details depend on your current M365 and Power Platform agreements. Speak with your Microsoft partner for specifics relevant to your organisation.
The self-healing desktop flows sound significant. How does it work?
When a desktop flow encounters a UI element it can no longer locate because an underlying application has been updated, the self-healing capability attempts to identify the correct element based on context and adapt the flow automatically. It’s not infallible, but it dramatically reduces the number of flows that break silently after system updates.
We’re not currently using Copilot Studio. Is this wave a good time to start?
The integration improvements in this wave, particularly the connection to M365 Copilot through Agent Builder and Work IQ, make Copilot Studio more accessible than ever for organisations already in the Microsoft ecosystem. If you’ve been considering it, this is a good moment to evaluate.
How do the new governance features help with cost management?
Granular Copilot credit tracking lets you see exactly who is consuming what and set caps before budgets are exceeded. Automated licence reclaim identifies and recovers unused licences automatically. Together, these give finance and IT teams the visibility and control they need to manage Power Platform costs at scale without manual monitoring.
Some of our flows break regularly when applications update. Will self-healing fix this?
Self-healing is specifically designed for this scenario. It won’t catch every possible change, particularly major application redesigns, but for the common case of minor UI updates causing flow failures, it’s a significant improvement. We’d recommend reviewing your most fragile flows and assessing which ones would benefit most from the update.