2026 Update: Building a Project Management Solution Using Plans in Power Apps

You watched Paul Stork build a complete project management solution using the Power Apps Plan Designer in early 2025. The platform has evolved rapidly since that recording, rendering many of those specific interface steps obsolete. Microsoft has shifted from standalone generative tools to a fully integrated, AI-native environment known as the "Vibe" experience.

Today, you do not just generate code; you collaborate with a digital software team of specialized AI agents. These agents handle everything from process mapping to complex data entry via the new Model Context Protocol (MCP).The licensing landscape has also undergone a massive shift with the official retirement of the Per App plan for new customers.

You saw Paul build a solution step-by-step using a single Copilot interface. Today, you would orchestrate the architecture through specialized agents, strict managed environments, and deeper Microsoft 365 integrations. This report provides the exact blueprint for building that same project management solution using the March 2026 toolset.

What's Changed Since This Session

The core methodology remains solving a business problem through relational data and low-code apps. However, the technical execution, governance, and licensing are entirely different in 2026.

How to Build This Today

The session showed Paul executing a highly manual workflow, interspersed with basic generative AI prompts, to stitch together his project management solution. The current ecosystem demands a more architectural approach. You are no longer just a maker; you are an orchestrator of AI agents.

We will walk through each of Paul's original demonstrations. For every scenario, I will outline the precise 2026 approach, detailing the new menus, configurations, and governance policies you must navigate.

Prerequisites Setup

The session showed Paul verifying Dataverse environments, assigning System Admin or Environment Maker roles, and checking Azure AI region availability. Today, you would build that same foundation, but with updated region support, stricter Managed Environment enforcement, and new administrative URLs.

First, navigate to thePower Platform Admin Centerto configure your foundational environment.You must ensure your environment has a Dataverse database provisioned. Without Dataverse, you cannot utilize the advanced multi-agent architecture or the Model Context Protocol.

Navigate through the left menu to Environments, select your target environment, and click Settings. Under the Users + permissions section, select Users. You must assign the System Administrator or Environment Maker security role to yourself and any co-developers.

Next, you must verify your region for generative AI features. In Paul's session, many non-US makers had to manually enable cross-region data movement. Today, Microsoft Foundry and Azure AI infrastructure have expanded significantly.

If you are located in the UK, Australia, or India, Copilot GPUs are now supported locally.You no longer need to route your data to the United States or Europe. This drastically simplifies compliance for government or strictly regulated project management solutions.

To verify this, stay in the Settings menu of your environment. Expand the Product category and click on Features. Scroll down to the Copilot and AI Builder section.

Ensure that the AI Builder preview models setting is enabled.If your region is natively supported, Copilot is enabled by default.If you are in a tier-two region, you must manually opt-out of cross-geo restrictions.

You must also verify your Dataverse storage capacity before beginning the build. The 2026 architecture generates significantly more metadata due to agentic logging. Navigate to Resources > Capacity in the admin center. Ensure you have at least 1 GB of database capacity available, or your pay-as-you-go meters will trigger overage charges.

Describe the Business Problem

The session showed Paul starting a prompt in the Plan Designer to build a simple project management solution with tasks, approvals, emails, client views, and dashboards. Today, you would build that same thing using the new AI-native workspace, treating the interface as a conversation with a digital software team.

You will not start in the traditional Power Apps portal. Navigate directly tovibe.powerapps.com.This is the new centralized, AI-native hub for solution generation.

The Vibe experience combines plans, data models, and apps into a single design surface.This minimizes context switching and keeps you focused on the business architecture. Click on the Create a plan button located prominently on the main dashboard.

A conversational interface will open. You do not need to write technical specifications. Paste your core business problem directly into the chat: "Build a simple project management solution to track tasks, manage manager approvals, send automated emails, provide external client views, and display management dashboards."

This interaction is fundamentally different from 2025. You are not generating a one-off app; you are initiating the Microsoft 365 Frontier Suite's agentic capabilities.The platform will analyze your prompt using multi-model intelligence to establish the foundation of your solution.

As soon as you submit the prompt, the system transitions from a passive interface into an active workflow participant.You will see the agent indicators spin up as they begin to parse your intent into actionable data structures.

Define and Edit Requirements, User Stories, and Roles

The session showed Paul manually tweaking generated user stories and defining security roles for project managers and team members. Today, you would execute this using the Requirement Agent, collaborating in real-time to shape the solution's parameters.

Once you submit your initial business problem, Vibe's Requirement Agent takes command.It will automatically parse your prompt and generate a structured, comprehensive list of user personas. For our project management scenario, it will typically isolate roles such as Project Manager, Client, and Team Member.

It will also output formal Agile user stories aligned with those personas. These are displayed in the central design surface, providing a clear text-based blueprint of the application's intended functionality.

Review these stories carefully. In the 2025 session, editing these required navigating through clunky configuration panes. Today, if a requirement is missing, you simply type a natural language command into the side panel chat.

Type: "Add a requirement for the Project Manager to reassign tasks to different Team Members." The Requirement Agent will dynamically update the user stories on the screen and adjust the associated persona permissions on the backend.

You can also refine non-functional requirements. Type: "Ensure all task data is highly secure and compliant with internal audit standards." The agent will log this parameter, which influences how the downstream agents configure the Dataverse security roles and environment routing.

Once the user stories accurately reflect your business needs, click the Approve Requirements button located at the bottom of the workspace. This action explicitly locks in this phase of the architecture. It seamlessly packages the context and hands it over to the next specialized AI agent in the workflow.

This handoff is critical. By forcing you to approve the requirements first, the platform ensures that no code or data models are generated based on ambiguous or incomplete instructions. This structured control plane drastically reduces technical debt.

Map Processes with Diagrams

The session showed Paul visually mapping out processes and decision trees for task approvals. Today, you would automate this visualization using the Process Mapping Agent, turning abstract requirements into concrete operational flows.

Within the Vibe workspace, the Process Mapping Agent will automatically analyze your approved requirements.It initiates an AI analysis wizard that guides the process step-by-step through the business problem, user processes, and proposed technology stack.

It then generates a dynamic, modifiable process diagram. This visual blueprint provides an end-to-end view of the process, detailing exactly how data flows through the application.It provides both a micro and macro vision of your solution, clearly illustrating activities, events, and decision points.

You will see a visual flowchart detailing the manager approval gate. To edit this logic, simply click on the specific node in the visual map. A properties panel will open on the right side of the screen where you can adjust the underlying logic.

For example, click the "Manager Approval" node. In the properties panel, change the logic using natural language: "Require secondary director approval for any tasks estimated over 40 hours."

The visual map will instantly redraw itself. It will add a new decision branch, insert a new persona for the Director, and reroute the approval flow lines. This real-time visual feedback ensures technical and business teams remain perfectly aligned.

This process map is not just a pretty picture; it is the actual architecture of your future Power Automate flows. By mapping out activities and decision points in advance, you identify bottlenecks early and ensure that the final generated solution matches the operational reality.

Once the flowchart perfectly represents your project management lifecycle, click Confirm Process Map. The digital software team will lock this blueprint and pass the context down to the data modeling layer.

Model Data

The session showed Paul adding relationships, columns, and users in the Plan Designer to establish the Dataverse schema. Today, you would orchestrate this using the Data Agent, leveraging advanced parsing tools to build complex relational models instantly.

Click on the Data Workspace tab within your Vibe plan.The Data Agent will have already processed your approved requirements and process map.It presents a complete Entity Relationship Diagram (ERD) on the canvas.

For this project management solution, it will suggest core tables such as 'Projects', 'Tasks', 'Approvals', and 'Clients'. It automatically defines the primary keys, foreign keys, and relationship behaviors (like cascading deletes) based on best practices.

Review the proposed columns within the ERD. In 2025, if you needed a new data point, you had to manually build it through the table designer UI. Today, do not build it manually. Chat directly with the Data Agent in the side panel.

Type: "Add a choice column to the Tasks table for 'Priority' with the options High, Medium, and Low. Make High the default value."

The Data Agent will immediately create the column, define the global choice list, establish the relationships, and apply the correct Dataverse data types. It handles all the schema generation in the background.

If your project management app requires offline capabilities for field workers, the data modeling phase now includes native advanced querying. You can configure offline data profiles directly within the studio.

Navigate to the offline profile editor for your app. Select the Filter option for your Tasks entity. Under the custom filter settings, choose Edit FetchXML.

The new FetchXML editor opens directly inside the Power Apps Studio.You no longer have to switch tools or rely on the limited Expression Builder. You can write precise query logic—such as filtering tasks by the current user and an 'Active' status—ensuring frontline workers only download the exact data they need.

This precise control over data filters improves offline performance and makes the application highly maintainable.Once your schema and queries are perfect, click Confirm Data Model to finalize the backend architecture.

Generate Artifacts

The session showed Paul generating Dataverse tables, canvas apps, model-driven apps, cloud flows, Power Pages sites, and Power BI reports from his plan. Today, you would finalize this step using the Solution Agent, outputting a modern, enterprise-ready package.

With your requirements, processes, and data modeled and approved, locate the Generate Solution button in the Vibe workspace.Click it. The Solution Agent now orchestrates the creation of all physical assets simultaneously.

The agent translates your abstract plan into a cohesive, deployable solution package. It provisions the Dataverse tables and establishes the relational links. It generates a robust model-driven app tailored for the Project Managers to handle heavy data entry and administration.

It also generates a highly responsive Canvas app designed for the Team Members. This app automatically utilizes the new Card modern control.This layout-aware control presents structured information—like task summaries—in a clean, responsive format that adapts seamlessly to both vertical mobile screens and horizontal desktop monitors.

For your external stakeholders, the agent generates a secure Power Pages site based on the 'Client' persona requirements you defined earlier.It wires up the Power Automate flows mapped out in your process diagram, ensuring automated emails trigger correctly upon task completion.

Crucially, because this is the 2026 update, the Solution Agent automatically embeds custom Copilot Studio Agents directly into your generated apps.It builds a 'Solution Agent' that understands the context of your specific project management data, ready to assist users from day one.

You will see the generated artifacts populate in your solution view. They are no longer isolated components; they are a unified, interconnected system built on a foundation of validated business logic.

Review and Customize Generated Artifacts

The session showed Paul refining flows and apps with Copilot, manually tweaking screens and logic. Today, you would customize these using enhanced embedded agentic features, modern design tools, and strict platform governance.

To customize your new apps, open the generated Model-Driven app in the designer. You will immediately notice a massive architectural shift: M365 Copilot chat is now natively embedded inside the application.

You no longer need to switch tabs or build custom side-panels. Users can ask natural language questions and reason over in-app project data seamlessly.As a maker, you can interact with this embedded Copilot to further customize forms or views without leaving the context of the app.

If your agents are performing automated tasks—like parsing incoming client emails and creating tasks—you can monitor them using the new Enhanced Agent Feed.This provides a shared workspace where human project managers can supervise agent activities. You can view side-by-side comparisons of agent-drafted records and approve them before they are committed to the database.

If you need to tweak the visual branding of your generated Canvas app, the process is no longer manual. You utilize the newly generally available Theme Copy-Paste feature.

Navigate to an existing, brand-approved corporate application. Copy the app's theme—which extracts all colors, typography, and styling tokens as a block of YAML code.Return to your new project management Canvas app, open the theme settings, and paste the YAML text.

This instantly standardizes your branding across the entire app portfolio, reducing repetitive setup and enforcing design governance at scale.You can even edit the YAML text manually if you need to make granular adjustments to specific hex codes.

While customizing the Canvas app, you should replace any clunky custom pop-ups with the new native Confirm() function.Navigate to a critical action button, such as "Delete Task". In the OnSelect property, wrap your patch function in a Confirm statement.

This will display a modal confirmation dialog that aligns perfectly with Fluent design principles.If the user clicks outside the dialog, the system safely treats it as a "no action" event, returning a blank value and preventing accidental data loss.

If you have pro-code developers on your team, the 2026 ecosystem allows them to extend the solution securely. They can build custom web components using React or Vue and deploy them directly into the Power Apps environment. These Code Apps are fully governed platform assets, seamlessly bridging the gap between low-code makers and professional engineering teams.

The SharePoint List Workaround (Non-Dataverse)

In Paul's session, he demonstrated a workaround for organizations that refused to adopt Premium licensing, using SharePoint lists as the data backend. While highly discouraged in 2026, this workaround is still possible. However, the process utilizes the massive New SharePoint Experience rolled out between March and May 2026.

Navigate to your corporate SharePoint site. You will see the completely redesigned interface. Click on the new Build section located on the modernized left-hand app bar.This is the new central hub for creating sites, libraries, and lists.

Select Lists and create your backend tables (e.g., 'Projects', 'Tasks'). Build out your columns just as you would in Dataverse, though you are limited to basic SharePoint data types.

In the past, you would have to navigate away to the Power Apps portal to connect this data. Today, the integration is seamless. Click the Integrate option directly on the top command bar of your Microsoft List.

Select Power Apps from the dropdown menu, and then click Create an app.Power Apps will automatically generate a basic three-screen Canvas app connected to your SharePoint list.

If you previously built standalone apps in the default environment using this workaround, be aware of new governance policies. Administrators are now aggressively utilizing the Power Platform advisor to migrate canvas apps and custom SharePoint forms out of the default environment.

Administrators use the Power Platform for Admin v2 Connector to automatically migrate these assets into designated managed environments.During this migration, they can quarantine or delete unmanaged apps. If you build a SharePoint workaround, you must coordinate with your IT team to ensure it is deployed to a governed environment, or risk losing your application entirely.

Licensing Quick Reference

The licensing model has simplified visually, but the financial enforcement is stricter than ever. The legacy "Per App" tier is officially dead as of January 2, 2026.You can no longer rely on chaotic, unenforced per-app metrics.

To deploy the advanced, AI-agentic solution recommended in this guide, your users will need the appropriate premium assignments. Here is the exact licensing structure you must procure to deploy this solution today.

If you are an existing Enterprise customer utilizing the deprecated Per App SKU, you will experience no immediate disruption. You can continue to use and renew this SKU until your current agreement ends.During your annual true-up process, you can adjust user counts.However, once your enterprise agreement expires, you will have a 60-day window to migrate your entire user base to an alternative solution, such as the Premium tier or the Pay-as-you-go meters.

The pay-as-you-go meters provide a highly flexible alternative for team members who only occasionally log in to update a task. By linking your environment to an Azure subscription, you are billed exactly $10 per active user, per app, per month.An active user is defined as someone who opens the app at least once in that calendar month; repeat access incurs no additional charges.

If your solution utilizes premium automated workflows—such as complex desktop flows or hosted robotic process automation (RPA)—these are also billed on a pay-as-you-go basis through Azure. Standard cloud flow runs cost $0.60 per run, while unattended desktop flows cost $3.00 per run.You must monitor these meters closely in your Azure portal to prevent unexpected cost overruns generated by highly active Copilot agents.