• Customer Self‑Service Portal
• Partner Portal
• Employee Self‑Service Portal
• Community Portal

When you provision a new site using any of these templates, it’s automatically built on the Enhanced Data Model with Bootstrap 5. No additional configuration is needed.

What’s new

• Enhanced Data Model (EDM) for key D365 templates is now generally available. New sites created with any of the four supported templates are automatically provisioned on the EDM, giving you the same modern architecture that Power Pages starter templates already use.
• Bootstrap 5 out of the box – All four templates have been migrated from Bootstrap 3 to Bootstrap 5, delivering improved responsiveness, accessibility, and alignment with current Power Pages UI standards.
• Full Design Studio and Management App support – The Power Pages Design Studio and Power Pages Management app fully support the new EDM‑based templates, including all updated attributes, relationships, forms, and views. Your authoring and customization experience remains seamless.

Why this matters

• Faster provisioning – Sites on EDM provision significantly faster, reducing the time from template selection to a live development environment.
• Streamlined ALM – Website configurations are solution-aware and stored in Dataverse solutions, with support for Environment Variables and Power Pipelines for deployments – making it easier to transport customizations across environments using standard application lifecycle management processes.
• Faster platform updates – Enhancements and bug fixes are delivered more efficiently on the Enhanced Data Model. Customers are not required to actively manage Power Pages solutions to stay up to date, feature updates and security patches are automatically shipped to ensure ongoing security and compliance.
• Modern, accessible UI – Bootstrap 5 brings a responsive grid system, improved accessibility defaults, and a cleaner component library, helping you deliver a polished experience to customers, employees, and partners without additional front‑end work.

With these four templates now on EDM, Dynamics 365 portals benefit from the same platform‑level investments and future enhancements as all other Power Pages sites.

Important: This update applies to newly provisioned sites only. Existing sites on the standard data model continue to operate without impact. Migration tooling for these four Dynamics 365 templates is not yet available but is coming soon.

How to get started

Prerequisites

• Sign in to the Power Platform admin center.
• Go to Manage > Environments > Select your Environment > Resources > Power Pages sites and ensure the Switch to enhanced data model toggle is turned on for your environment.
• Verify your environment has Power Pages Core Package v1.1.2602.230 or later. For upgrade steps, see Update the Power Pages solution.

Create a new site

1. Open the Power Pages home page.
2. Select Create a site.
3. Choose one of the four supported Dynamics 365 templates and select ‘Choose this template’.
4. Fill in the required details and select ‘Done’.
5. Your new site is provisioned on EDM with Bootstrap 5 automatically.

Learn more

• Enhanced Data Model overview – https://learn.microsoft.com/en-us/power-pages/admin/enhanced-data-model
• Power Pages Management app – https://learn.microsoft.com/en-us/power-pages/configure/portal-management-app

We’d love to hear your feedback! Share your thoughts on the Power Pages community forum.

The post Enhanced Data Model and Bootstrap 5 Now Available for Dynamics 365 Portal Templates appeared first on Microsoft Power Platform Blog.

First introduced in preview, Server logic was designed to simplify how makers and developers implement server-side operations. With GA, Server logic is now ready for enterprise production workloads, backed by strengthened governance, improved extensibility, and the reliability customers expect from Power Platform.

Alongside GA, we’re also announcing two important enhancements that further reinforce enterprise readiness:

• Governance control to disable external calls
• Support for unbound Dataverse custom actions

Governance control to disable external calls

Security and compliance are critical considerations for enterprise applications. Many organizations require strict control over outbound connectivity to ensure adherence to internal policies and regulatory requirements.

With GA, administrators can now disable external calls from Server logic layer.

Support for unbound Dataverse custom actions

Extensibility is key to building scalable and reusable solutions. With GA, Server logic now supports unbound Dataverse custom actions, enabling deeper integration with your existing business logic layer.

Get started

Server logic is now available for production use in Power Pages. Whether you’re enhancing an existing site or building a new application, you can take advantage of its capabilities to deliver more intelligent, secure, and governed experiences.

• Server logic overview
• Server objects
• Tutorials
• Disable external service calls from server logic

The post Announcing General Availability (GA) of Server Logic in Power Pages appeared first on Microsoft Power Platform Blog.

In a recent conversation I had with Dion Hinchcliffe at Futurum, we spent time unpacking a shift I’m seeing consistently across enterprises experimenting with AI. It’s not just about copilots or chat interfaces. It’s about something deeper: a change in how work is designed, governed, and operated when systems can reason and act with intent.

For decades, applications have been the primary interface between people and systems. Work meant navigating menus, filling out forms, and clicking through screens carefully designed to constrain what users could do. Productivity improvements came incrementally—better layouts, faster load times, and more automation behind the scenes—but the underlying engagement model stayed the same. People adapted to software.

That model no longer holds.

As organizations race to adopt AI, a new challenge is becoming clear: translating human intent into systems that can act autonomously—without sacrificing control, security, or trust. Intent-first development addresses that gap by reshaping how agentic applications are designed, governed, and delivered at scale.

Agents as the new interaction layer

Instead of teaching people how to use systems, we can let people express intent—and allow systems to determine how that intent is carried out. This is not about replacing all apps overnight. It’s about changing their role. Apps no longer need to expose every possible action through UI. Instead, they:

• Provide trusted capabilities the agent can invoke
• Enforce business rules and permissions
• Act as systems of record, not systems of navigation

As AI systems become capable of reasoning, acting, and adapting, organizations are beginning to rethink the relationship between humans and software. In an agentic model, the agent becomes the primary interaction surface. A user may no longer need to know which system to open or which workflow to follow. They can simply state what they want to achieve: open a purchase order (PO), resolve this case, prepare a customer briefing.

Behind the scenes, agents orchestrate the necessary steps across systems, policies, and data sources. Procurement rules are applied. Approvals are routed. Records are updated. The user expresses intent once; the system coordinates the work.

Agentic solutions aren’t eliminating applications, but they are changing how people engage with them. Apps are the trusted capabilities agents rely on—serving as systems of record, sources of authority, and enforcement points for business rules and permissions. Applications shift from user destinations to services agents invoke. Agents work because structure already exists.

Rethinking enterprise complexity: Orchestration over navigation

This shift becomes clearer when you look at everyday enterprise processes.

Take something as common as opening a purchase order. Today, that often means navigating multiple tools, involving several teams, and manually coordinating approvals. The complexity isn’t the work itself—it’s knowing how to move through the systems.

With an agent‑first approach, that complexity is inverted. A user can simply say they need to open a PO for a project. The agent determines which background agents are required—vendor management, policy validation, approvals—and orchestrates the process across systems without forcing the user to navigate them.

We see the same pattern emerging in CRM. Rather than sales teams manually updating records, agents can monitor emails, calls, calendars, and systems in the background—keeping data current and surfacing relevant context proactively. The agent becomes the interface to customer intelligence, while the CRM remains the authoritative store behind it.

The value here isn’t conversational UI for its own sake. It’s reducing cognitive load while preserving control.

Agents as the business logic and decision layer 

This shift also changes where business logic lives.

Traditional enterprise systems embed logic deep inside individual applications—rules, workflows, and decision trees hardcoded into each tool. That makes change expensive and reuse difficult. When requirements evolve, logic must be rewritten repeatedly across systems.

Agentic systems invert that model. Logic moves into a shared reasoning layer that sits above systems of record. Agents evaluate intent, context, and constraints, then determine which actions are required right now. Policies, best practices, and exceptions can be defined once and applied consistently across processes instead of being repeatedly embedded in individual applications.

This is where the economics of software start to change. Improvements to reasoning or decision quality can compound across organizational functions—HR, finance, operations, and customer engagement—without rebuilding each system individually. Business value shifts from static workflows to shared enterprise intelligence.

Headless agents as a new layer of digital labor 

Not all agents interact directly with people.

Many of the most impactful agents operate quietly in the background—monitoring systems, reacting to triggers, coordinating tasks autonomously. These “headless” agents update records, flag issues, generate reports, and escalate decisions only when human judgment is required.

Together, conversational and headless agents form a new layer of digital labor. Routine work is handled automatically. Humans stay focused on oversight, judgment, and exceptions. The agent doesn’t replace enterprise logic—it coordinates it.

Operating agentic systems at scale requires a control plane

One point Dion and I kept coming back to is this: the real challenge with agentic systems isn’t building the first one. It’s operating hundreds—or thousands—of them responsibly.

As agents scale across teams and geographies, the questions shift quickly. How do you maintain visibility into what agents are doing and why? How do you enforce security, policy, and compliance consistently as agents act across systems? How do you measure impact, cost, and effectiveness as usage grows?

Without a managed platform, intent first development becomes ungovernable at scale. Logic fragments. Visibility breaks down. Early experimentation turns into operational risk. Governance must mature alongside autonomy.

This is where enterprise readiness becomes decisive.

Governance, lifecycle management, observability, and control aren’t optional add‑ons. They’re the foundation that allows agents to operate safely and reliably. Successful enterprise adoptions hide complexity behind an interface that works the way people already think.  Agents don’t eliminate the need for structure—they depend on stronger, more explicit structure than traditional automation ever required.

From pilots to an enterprise operating model

Most organizations begin with pilots—and that’s the right place to start. But pilots stall when governance, ownership, and measurement are treated as afterthoughts.

The pilots that scale share common patterns: centralized policy management, clear accountability between IT and business teams, built-in monitoring, and an explicit path from experimentation to production. Governance isn’t what slows progress; it’s what gives leaders confidence to move faster.

Over time, this becomes more than a collection of use cases. It becomes an operating model. Work shifts from task execution to outcome driven orchestration. Processes move from periodic redesign to continuous optimization. Systems adapt as business intent evolves.

Building adaptive enterprise systems for an agent-first world

This shift isn’t about predicting the future. It’s about building systems that can adapt as it arrives.

Agentic transformation isn’t just a technical change. It’s an operational one—reshaping how work is designed, governed, and continuously improved across the enterprise. Organizations that invest early in the right foundations—clear intent, strong constraints, and disciplined scale—will be positioned to turn intelligent applications into a durable advantage, not a fleeting experiment.

The most successful organizations won’t ask how to bolt agents onto existing apps. They’ll ask how to redesign systems so agents can sit confidently at the front door—turning intent into action with trust, speed, and scale.

In an agent first world, applications remain systems of authority and agents simply coordinate how and when those capabilities are invoked. Apps evolve:

• From destinations → to services
• From user driven workflows → to agent orchestrated actions
• From “where work happens” → to “how work is made possible”

If you want to hear this thinking unpacked in more detail, I explore these ideas directly with Dion Hinchcliffe at Futurum—from agents as the new interaction layer, to why governance becomes more critical, not less, as autonomy increases. Our conversation gets into real enterprise examples, the challenges of moving beyond pilots, and what it actually takes to operate agentic systems at scale.

I encourage you to watch the full interview to hear how these concepts show up in practice and to learn how intent first development is shaping the future of enterprise AI.

The post From apps to agents: Rearchitecting enterprise work around intent appeared first on Microsoft Power Platform Blog.

Summary Welcome to the Power Platform monthly feature update! We will use this blog to share news in Power Platform from the last month, so you can find a a summary of product, community, and learning updates from Power Platform in one easy place. Now, let’s dive into what’s new in Power Platform:

Apps, agents and Copilot

Public preview: M365 Copilot chat in model-driven apps

Copilot chat is now available directly inside apps built with Power Apps, bringing the intelligence of Microsoft 365 Copilot into the flow of business processes.

This unified experience—currently limited to model-driven apps—lets users ask questions, reason over in‑app data, and connect insights from documents, communications, and collaboration—without leaving the application they’re working in. By embedding Copilot chat into low‑code apps, organizations can keep users in context and in flow, reducing app switching while accelerating decision‑making. Teams can also leverage powerful first‑party agents like Researcher and Analyst, as well as custom Copilot Studio agents, to analyze data, generate insights, and take informed action directly within their apps.

To manage Microsoft 365 Copilot chat for model-driven apps, start by learning how to manage Microsoft 365 Copilot chat. Power Platform administrators can set up and configure the Microsoft 365 Copilot chat feature for users in their environment and makers can then enable or disable Microsoft 365 Copilot chat for a specific model-driven app.

Public preview: Power Apps MCP and enhanced agent feed

We’re bringing Power Apps Model Context Protocol (MCP) Server and an enhanced agent feed into public preview. This is a step to enable better human-agent collaboration directly inside business applications with built‑in human supervision.

Power Apps MCP brings agentic features from apps to agents as tools – starting with data entry. Agents will be able to parse the unstructured data into forms that users use in apps and create records directly, as well as flag them for human review or action.

The enhanced agent feed provides a shared workspace for humans to oversee the agent activity—makers can provide granular visibility into agent actions for their users, use side‑by‑side comparison views for approvals, and direct navigation to in‑app records.

Building modern apps

Public preview: a new modern Card control

This new modern Card control helps makers build clean, responsive, and consistent UI layouts in canvas apps.

The modern Card control allows makers to present structured information—such as summaries, previews, and tiles—using a single layout‑aware control instead of composing multiple classic controls. Cards automatically adapt to vertical or horizontal layouts and align with Fluent UI design principles, improving visual consistency across apps.

By reducing layout complexity and improving responsiveness out of the box, the Card control enables faster UI composition while supporting accessibility and scalability across screen sizes.

Generally available: theme copy‑paste

With theme copy-paste it is it easy to reuse visual styles across canvas apps without manual reconfiguration.

Theme copy‑paste allows makers to copy and reuse a Canvas app’s theme—including colors, typography, and styling tokens—across other apps. These themes are copied as YAML which can also be edited manually by makers as text. This reduces repetitive setup and helps ensure consistent branding and visual identity across an app portfolio. 

By simplifying theme reuse, this update accelerates new app creation and supports design governance at scale, especially for teams managing multiple canvas apps across environments.

Generally available: confirm() function in canvas apps as a fluent dialog

The Confirm function displays a modal confirmation dialog over the current canvas screen, prompting the user to explicitly confirm or cancel before continuing. In canvas apps, there’s also a dismissal path (for example, clicking outside the dialog) that is treated as no action and returns blank.

The canvas experience is designed to align with Fluent dialog behavior and to respect the current app theme. You need to have modern controls turned on to get fluent dialog, else you will get a browser native dialog.

Managed platform

Public preview: move canvas apps and custom SharePoint forms out of the default environment

The default environment in Power Platform often becomes a shared space where makers build and test applications, leading to potential challenges with governance and organization. As resources accumulate in this environment without structured oversight, administrators may face difficulties managing security policies, tracking ownership, and maintaining compliance across their tenant.

We introduced a recommendation in Power Platform advisor, as a preview, that enables administrators to migrate certain canvas apps and custom SharePoint forms from the default environment to designated managed environments. The migration can be done manually from the Recommendations page under the Actions menu in the Power Platform admin center or automated using the Power Platform for Admin v2 Connector. When moving apps, administrators can choose to keep the original resource as is or restrict access to it by quarantining, or deleting it entirely.

This helps Power Platform administrators in implementing effective governance and DLP controls and establish clearer boundaries for app development.

Generally available: host and run code apps in Power Apps

We’re excited to announce that code apps in Power Apps are now generally available, empowering developers and IT alike at a moment when organizations are building more custom applications than ever. With the rise of AI‑accelerated and code‑generation‑assisted development, teams can build high‑quality web apps faster than before while IT faces mounting expectations around governance, security, and operational oversight. Code apps bridge that gap by giving developers full code‑first flexibility and giving IT the enterprise‑grade guardrails needed to manage a growing app landscape

Power Apps code apps bring the full strength of Power Platform to web developers. Build with popular frameworks (React, Vue, and others) in any code-first IDE, and deploy to Power Apps. Every code app automatically becomes a governed Power Platform asset, giving IT visibility and control without creating friction for developers. 

Learning updates

Training paths and labs

Updated training

• Create and customize sites with Power Pages design studio – Training
• Explore Power Pages design studio data and security features – Training
• Explore Power Pages templates – Training

Power Apps maker

New

• Use preview tools and upcoming features in Dataverse MCP server
• Use modern themes in canvas apps  

Updated

• Create or edit model-driven app web resources to extend an app
• Generate a page using natural language
• Overview of pipelines in Power Platform
• Admin deployment page
• Use table segmentation in solutions
• Add Microsoft 365 Copilot chat for app user in model-driven apps
• Overview of the new Power Apps vibe experience
• Create apps, data, and plans together using Power Apps vibe
• Card modern control in Power Apps
• Keyboard shortcuts for canvas apps
• Sign in to Power Apps

Power Platform administration

New

• Move apps from the default environment (preview)

Updated

• Virtual Network support overview
• Set up virtual network support for Power Platform
• Automatic deletion of Power Platform environments
• Microsoft 365 environment for recurring Copilot prompts
• Preview and experimental AI models
• Export data to Application Insights
• General availability deployment
• PowerShell support for Power Apps and Power Automate
• Environment routing
• Manage Copilot Studio credits and capacity
• Buy Power Automate licenses
• Restore deleted Microsoft Dataverse table records (preview)
• Dataverse capacity-based storage details
• Configure Dataverse search for your environment
• Track, manage, and scale Copilot adoption in the Power Platform
• Content security policy
• Continuous access evaluation (preview)
• Power Platform inventory
• Finance and operations storage capacity
• Permissions and roles for SharePoint document management tasks
• Allow external language models for generative responses
• View and manage on-premises data gateways
• System Settings Service tab
• Administrator analytics and reports for Microsoft Power Apps
• Microsoft Power Platform admin documentation
• Power Platform inventory API
• Discover what drives engagement in the Power Platform by using the Usage page
• Configure email resolution with the DoNotResolve flag
• Licensing and AI Builder credits
• Licensing and Copilot Credits
• End of AI Builder credits
• Workflows agent privacy and data subject request (DSR) FAQ

Power Platform developer

New

• Web API Query Data Sample (PowerShell)
• Web API Conditional Operations Sample (PowerShell)
• Web API Functions and Actions Sample (PowerShell)
• dataApi reference for generative pages
• RetrieveTimeZone sample (C#)

Updated

• Power Apps code apps overview (preview)
• How to: Connect your code app to data (preview)
• How to: Connect your code app to Dataverse (preview)
• navigateTo (Client API reference) in model-driven apps
• How to: Configure Content Security Policy (preview)
• Generate a page using natural language with model-driven apps in Power Apps
• Microsoft Power Platform CLI auth command group
• Dataverse SDK API Reference
• Microsoft Power Platform API reference

Power Apps user and mobile

Updated

• How mobile offline works in Power Apps
• Use Microsoft 365 Copilot chat in model-driven apps (preview)
• Optimize the offline profile
• Set up mobile offline for canvas apps
• FAQ for Dataverse search
• What is Dataverse search?
• Visualize data in a view with Copilot (preview)
• Customize and build your mobile app using the wrap wizard
• Prevent canvas app restarts in the Power Apps mobile app
• Check user access to a row
• Use Copilot’s form fill assistance feature in model-driven apps
• Use Copilot chat in model-driven apps
• Modern, refreshed look for model-driven apps
• Supervise agents in model-driven apps with agent feed

Power Pages

New

• Connect Power Pages server logic to Azure Function
• Power Pages Server Logic Debugging Guide

Updated

• Create an event registration page
• How to interact with Dataverse tables using server logic
• Power Pages server logic overview (preview)
• Advanced Settings (preview)
• Create and deploy a single-page application in Power Pages
• Set page permissions
• Security agent (preview)
• Add Copilot summarization to case page (Preview)
• Configure Omnichannel with Power Pages site agent
• Edit code with Visual Studio Code for the Web (preview)
• Use the Visual Studio Code extension
• FAQ for site search with generative AI
• Set up Microsoft Entra External ID with Power Pages

The post What’s new in Power Platform: February 2026 feature update appeared first on Microsoft Power Platform Blog.

Bringing apps and agents together for organizational transformation driven by human-agent collaboration

Today, we’re excited to bring the Power Apps MCP (Model Context Protocol) Server to public preview. The Power Apps MCP Server enables agents to automate repetitive app tasks, with human review and approval through the new task-centric enhanced agent feed. The Power Apps MCP Server also provides built‑in supervision tools that allow humans to review, assist, and take control of agent actions when needed.

Building on success: The journey from form fill assistance to MCP server

We introduced the agentic data entry in Power Apps some time back to speed up form completion with AI suggestions. The response from users was overwhelmingly positive, with high adoption and valuable feedback that shaped our vision for what came next.

Based on this feedback, we recognized an opportunity to extend these capabilities beyond a single-purpose agent. The result is the Power Apps MCP Server – the same proven technology that powers the in-app form fill suggestions – now available as a flexible MCP tool that can be used by agents across your organization and scale to meet enterprise demands. We’re starting with data entry and will be adding other create-read-update-delete operations over time, progressively expanding the app tools available to your agents.

The Power Apps MCP Server equips your agent with two types of capabilities:

1. Automate repetitive app tasks: agents can now use tools like form filling assistance initially built and refined in Power Apps. Just add and configure the MCP tool in your agent and point the invoke_data_entry tool to a shared mailbox, SharePoint folder, or other unstructured content source. The invoke_data_entry tool will extract the relevant fields from the source and create the corresponding records in your app, with built-in human review and approval in the agent feed.

2. Supervise agent activity: agents can now handoff control to humans for review, assistance, and steering through these MCP tools. This way business users can supervise agent activity in the agent feed. The tools give makers much more control over the tasks they want to publish to the agent feed and when they need agent-human handoff.

Agent feed reimagined with native MCP integration

To leverage the capabilities of the Power Apps MCP Server, we’ve completely overhauled the agent feed. This transformation gives makers more granualar control to shape the agent feed experience – deciding which tasks appear and how users track and interact with agent activity. The redesigned agent feed provides a shared collaborative space where humans and agents work together seamlessly:

• Makers have granular control over agent feed tasks surfaced to users.
• Users get a side-by-side comparison view for data entry tasks that makes it easy to review and approve agent actions.
• Direct navigation to specific records within your app for context-aware interactions
• Users can view agent performance metrics and insights in their app.

Turn your existing apps into agentic apps

With the Power Apps MCP Server, makers can bring AI agents directly into the apps their users already know and use. This allows business experts to easily supervise, collaborate with, and guide agents as part of their daily workflows. As a result, both traditional and new business applications evolve into agentic apps, delivering rich, agent‑enabled experiences, see details here.

State Farm reimagines claims processing

State Farm’s team processes dozens of claims estimates every day, each arriving in different formats with varied attachments and information structures. Manually reviewing and entering this data is time‑consuming and prone to errors.

With MCP powered agents and agent feed, these emails can be automatically analyzed to extract key field values, and any missing details are surfaced for human review before the records are created. Learn more in the Microsoft Ignite 2025 session.

Learn more, get started and share your ideas

We invite you to learn more about agent feed in our learn documentation and explore what it can do for your apps. The Power Apps MCP server is being made available in public preview starting with United States early release cycle environments  and will be gradually rolled out to other regions, following standard weekly deployment.

We’d love to hear your experience and feedback – share it in Community forum or connect with us directly. Let us know what you are building, where supervision is useful and what would you welcome as the next iteration on the agent supervision capabilities in the apps.

The post Public preview: Power Apps MCP and enhanced agent feed for your business applications appeared first on Microsoft Power Platform Blog.

Since the launch of the preview, we’ve heard the feedback from developers who want to bring modern front-end frameworks to Power Pages while still benefiting from built-in authentication, Web API access, and Dataverse integration. Today’s GA release makes that experience production ready.

What’s new in this release

• Generally available: SPA support is now production-ready and fully supported, starting with site versions 9.8.1.x and later.
• Styling fixes: We’ve resolved issues where Power Pages platform styles could override your custom CSS.
• Improved documentation: Updated guidance for authentication configuration and local development setup.

Pro tips for SPA development

Disable profile page redirect for SPAs

By default, Power Pages redirects users to the built-in profile page (/profile) after login. For SPA website, this redirect can disrupt your app’s navigation flow. Disable it by adding this site setting:

After updating this setting, users are redirected to the returnUrl specified in your login form, which defaults to the current page or your app’s home page.

Set up local development

You can run your SPA locally with full authentication and Web API access to accelerate your development cycle. This means JavaScript hot reload, local debugging, and immediate feedback without deploying changes to your portal every time. For step-by-step instructions, see Set up local development for SPA sites.

Resources

• Create single-page application (SPA) for Power Pages
• Power Pages Web API overview

We are looking for your feedback

Your feedback helps us improve the developer experience on Power Pages. Share your thoughts and reach out on the Power Pages Community Forum. You can also submit ideas through the Power Pages Ideas portal.

The post Announcing General Availability (GA) of building single-page applications for Power Pages appeared first on Microsoft Power Platform Blog.

Designed specifically for model-driven apps, generative pages offers a new way to augment or supplement your existing applications with custom user experiences. You provide natural language instructions, and the agent generates a fully functional React page connected to your data, ready to fit seamlessly into your app.

What makes generative pages different?

Generative pages easily connect to your data in Dataverse, ensuring your pages are deeply integrated with your app. And because they support virtual entities, you can extend your app by connecting to external sources like SharePoint or Dynamics 365 Finance and Operations, without adding complexity.

These new AI-generated pages that are a natural part of your model driven app even adhere to the modern theme that is in use in your app. You also have full transparency: after each iteration, you can look at the code changes to see exactly what the agent updated in that iteration. And you have full control and can take the wheel to edit the code manually when you need. Pages are also solution-aware, so you can easily manage them across environments as part of your ALM process.

Imagination at the center

With Generative Pages, what you can build is limited only by your creativity. There are no rigid templates or layout constraints, allowing you to create visually rich, productive experiences for your end users.

Generative pages stands out as one of the most remarkable and significant innovations in Power Apps that I have encountered throughout my 5+ years of experience working with the Power Platform. It has facilitated collaboration among customers, designers, business analysts and developers, allowing them to jointly shape UI and UX within minutes! The possibilities are endless with regard to what types of pages you can create. I’m excited to see the feature evolve and see what pages our 2,500 makers create!”

–James Connelly, Power Platform Technical Lead, Network Rail

* Any data shown in the above pages relate to fictional data only and do not represent actual incidents or Network Rail employees

Get to a working solution fast

Using generative pages also simplifies the development cycle. What once took days, weeks, or even months to build using custom coding or meticulous drag-and-drop building, can now be started with a single prompt and finished in a fraction of the time, often within minutes or hours for a functional page, and just a day or two for a fully polished custom user experience that is part of a performant model driven app.

With Generative Pages in Power Apps, I was able to go from concept to a fully functional experience in just one day—something that previously required weeks of coordination between design and development teams. This is a game-changer for accelerating delivery and empowering makers.”

-Angel Limas, CIO, Power Platform, Product & Strategy + Demand Lead, Accenture

Get building with the generally available generative pages

Generative pages have been in public preview for almost 4 months. During this period, we’ve advanced the technology further and with the general availability announcement, you can now use generative pages with:

• Production support in U.S. environments
• Fluent UI controls with default alignment to your app’s theme
• Solution-aware components for seamless movement across environments
• Powered by GPT-5 for best-in-class code generation quality and performance

Unlock the new era of AI-driven app development and start building today. Experience how fast, flexible, and creative app development can be – log in or provision your Power Apps Developer Plan today.

The post Generative pages in Power Apps is now generally available appeared first on Microsoft Power Platform Blog.

Forget the days of jumping through hoops to call external APIs, orchestrate complex actions or securely handle sensitive tasks. Server logic places real power behind your pages

Why Server Logic?

Modern websites do more than display content. Customers expect transactions, workflows, decisions and trusted data flows. Server logic turns Power Pages into a backend-enabled experience.

With server logic, makers can:

• Connect to external services like ERP, CRM or banking systems
• Run business workflows without exposing secrets in front-end code
• Validate inputs and manipulate data on the server
• Trigger actions based on site visitor interactions
• Keep sensitive operations secure and compliant

How It Works

Server logic executes securely within the Power Pages infrastructure. Makers can author logic using JavaScript that run on the server-side whenever a page needs intelligence beyond what is available on the browser.

Key building blocks:

• Server scripts: Write logic to call APIs, transform data, run workflows or reuse platform connections and policies.
• Server objects: Use built-in modules that give you safe access to HTTP calls, Dataverse operations and more.
• Event triggers: Invoke logic based on form submissions, button clicks or programmatic calls from client side.

Refer the Server logic documentation and quick starts for more details.

Secure External Integrations

Calling external services has always been high stakes: secrets, authentication, network policies. Server logic makes it straightforward and safe.

Examples:

• Fetch loyalty points from a third-party rewards system
• Validate user identity with government or industry standard verification APIs
• Send loan application information to an internal banking engine
• Push form submissions into a custom Azure service

On the road to General Availability

We are continuously expanding server logic feature. Enhancements planned ahead of GA include:

• Support for unbound Dataverse custom APIs
• Availability in sites running on Standard Data Model
• Invocation from Liquid objects

• Ability to retrieve real time data from Dataverse
• Server Logic playground with debugging support

We are looking forward to your feedback

Your feedback is crucial in shaping the future of this feature. We want to hear from you!

The post Introducing Server logic in Power Pages: Business Logic Right Where It Belongs (Preview) appeared first on Microsoft Power Platform Blog.

Why Use GitHub Spark + Codespaces and Host in Power Pages?

GitHub Spark is an AI-powered tool that generates single-page applications using natural language prompts, while Codespaces provides a pre-configured development environment in the cloud.

Power Pages serves as a secure, scalable, and enterprise-grade platform tailored for modern web development. With built-in capabilities for security, governance, analytics, and seamless integration across the Microsoft Power Platform ecosystem, it provides everything you need to build, manage, and scale professional-grade web applications efficiently. Together, these enable:

• AI-assisted site generation with GitHub Spark
• Instant cloud-based coding with GitHub Codespaces
• Seamless deployment to Power Pages using Power Platform CLI
• Built-in security, governance, analytics, and seamless Power Platform integration delivered through Power Pages

Hosting Your Site in Power Pages

Once your site is created using Spark and edited in Codespaces, you can deploy it to Power Pages using the Power Platform CLI.

Quick steps:

1. Use GitHub Spark to generate your site template.
2. Open the project in GitHub Codespaces.
3. Customize your site using React or other SPA frameworks.
4. Deploy using pac pages upload-code-site.

To explore the complete step-by-step guide, refer to the document: Tutorial: Create and deploy a single-page application using GitHub Spark and Codespaces.

Demo

Key Resources

• Tutorial: Create and deploy a single-page application using GitHub Spark and Codespaces.
• Create single-page application in Power Pages
• Power Pages sample sites

The post Build and deploy a single-page application in Power Pages using GitHub Spark and Codespaces appeared first on Microsoft Power Platform Blog.

Aligning to the official GPT-5 announcement yesterday, makers using generative pages will have access to new GPT-5 model, which offers significant improvements in code generation quality, performance, and reliability. To give optimal flexibility, we’ve added a model selector to the experience—allowing you to choose between GPT-4.1 and GPT-5 depending on your needs. This update started rolling out yesterday.

What’s new with GPT-5 in generative pages?

• Improved code generation quality: Better syntax, smarter layout handling, and fewer rendering errors.
• Fewer failures: Higher success rates for generated pages, even in complex scenarios.
• Enhanced performance: Faster generation times during the page creation process.

In addition to giving access to the new GPT-5 model, this release includes several minor enhancements to improve usability and reliability:

• An expanded context window to reduce lost requirements in longer sessions
• Better fallback styling to preserve model-driven app design
• Ensured retention of image attachments during retries
• More informative error messages for invalid requests
• Conditional warnings when another user is editing the same page
• Improved undo/restore behavior for error and retry scenarios

This update reflects our commitment to making AI-native app development intuitive, flexible, and enterprise-ready.

Explore the documentation here to learn more and start building with GPT-5 today.

The post Generative pages get better with GPT-5: Improved quality, enhanced performance appeared first on Microsoft Power Platform Blog.