PL-900 · CHEAT SHEET

Microsoft Certified: Power Platform Fundamentals

Key concepts grouped by official exam weight.

Download as PDF (free, no sign-up)

1. Describe the business value of Microsoft Power Platform

5–10%

Describe the business value of Microsoft Power Platform services

Power Apps lets people build custom business apps with little or no code that run in a browser and on mobile devices and connect to Dataverse or to many online and on-premises sources. Canvas apps give full control of the layout; model-driven apps, built on Dataverse, derive much of the interface from the components you add.
Power Automate streamlines business processes and automates repetitive tasks with little or no code, from simple notifications to processes across several apps. Cloud flows start automatically, instantly or on a schedule and use connectors; desktop flows add RPA for tasks on the web or the desktop, including legacy applications.
Dataverse securely stores and manages the data used by business applications, organized in tables of rows and columns. It includes standard tables for common concepts and lets you create custom tables and import data with Power Query. Its metadata is used directly in Power Apps, and role-based security controls who can access each table.
A connector is a wrapper around an API that lets apps and flows communicate with a service. It can provide tables, actions or both, and many offer triggers that start flows. If no prebuilt connector exists, you can create a custom one, and the on-premises data gateway gives secure access to on-premises resources.
Power Pages is a secure, enterprise-grade, low-code SaaS platform for creating, hosting and administering external-facing business websites. It offers templates and a design studio, uses the same shared Dataverse data as other Power Platform components, and protects information with site visibility, web roles, and table and page permissions.

2. Manage the Microsoft Power Platform environment

20–25%

Describe Microsoft Dataverse

Dataverse stores information like a traditional database, but it is a SaaS service that also keeps business logic, forms, charts and dashboards with the data. It includes role, row and column security, backups and encryption, along with visual designers and rich data types. It can be combined with existing databases without migrating them immediately.
In Dataverse, a table holds rows and columns; it can be standard, activity, virtual or elastic, and its ownership cannot change after creation. Columns have a data type and appear on forms and in views. Relationships are 1:N, created by a lookup column, or N:N, using an intersect table, and their behavior decides what happens to related rows.
Forms are the interface for viewing and editing a row: main, quick create, quick view and card, each with its own purpose. Views define how a list of rows is displayed: columns, order, width and default filters. Views are personal, system or public; public views appear in the view selector and can be used in subgrids and dashboards.
Dataverse offers several business logic options: no-code business rules to validate data, Power Fx in formula columns and reusable server-side functions, business process flows that guide people through stages, and Power Automate to automate across applications. Power Fx is the low-code language of Power Platform, with a syntax similar to Excel formulas.
With Copilot, a maker describes the data in natural language and gets tables, columns and relationships that can be reviewed and refined through conversation. A maker can also start from an Excel file or SharePoint list, for which Copilot proposes names and data types. The generated output should be reviewed before the table is created, because it might not be fully accurate.

Describe Microsoft Power Platform administration and governance

An environment is a space to store, manage and share data, apps and flows, and a container that separates apps with different roles, security requirements or audiences. It belongs to a tenant and a location, and has zero or one Dataverse database. Its type indicates the purpose: permanent work (production), testing (sandbox) or experimentation (default).
Microsoft Entra ID authenticates users and, in Dataverse, security roles group cumulative privileges whose access level ranges from User to Organization. Business units and teams define which records people see, column-level security protects specific fields, and environments act as security boundaries, with security groups controlling who can be a member.
For privacy, Power Platform offers a choice of geographical location, encryption in transit and at rest, customer-managed keys, Customer Lockbox and steps for responding to Data Subject Rights requests. For accessibility, the Accessibility checker reviews canvas apps and Power Pages follows global standards, although whoever customizes a site is responsible for keeping that content accessible.
The Monitor area of the admin center measures the operational health of apps, flows and agents and suggests improvements, and analytics reports show the use of Dataverse, Power Apps and Power Automate. Makers have Live monitor for diagnosing sessions, the automation center for flows and the Monitor page in Copilot Studio for agents, while Application Insights extends the telemetry available.
Solutions are the ALM mechanism: they are developed as unmanaged and deployed as managed. Pipelines in Power Platform automate that move between environments with upfront validation, solution backups and ordered stages, without copying table data. Makers deploy from their development environment, while admins govern through a custom host, approvals and delegated deployments.

3. Demonstrate the capabilities of Power Apps

20–25%

Describe Power Apps capabilities and use cases

Canvas apps start from a blank surface where the maker places components and controls the whole layout without writing code. They connect to Dataverse and many other sources through connectors, can be built with Copilot, embedded in Teams or SharePoint, and run in a browser or on mobile. When you share one, you must also share its data sources.
Model-driven apps are built by adding components such as forms, views, charts and dashboards to Dataverse tables; without a data model in Dataverse you cannot create one. They suit complex, data-dense processes with related records. Components largely determine the layout, the apps are automatically responsive, and solutions move them between environments.
Plans in Power Apps (Plan designer) turn a natural-language description, with optional images, into a Power Platform solution. Three agents generate the user requirements, the data model and the technology proposal. Plans need a Dataverse database; you save the plan and tables in a solution before creating objects, and you can also document an existing solution.
Code apps let developers build business web apps in a code-first IDE, using frameworks such as React or Vue with full control of the UI, and run them in Power Platform. They provide Microsoft Entra authentication, access to connectors and adherence to managed platform policies. An admin must enable them on each environment.

Describe how to build basic apps by using Power Apps

Copilot in Power Apps creates Dataverse tables from a description of the data to track, and makers can refine them in natural language before building the canvas app. A plan can also start a data-connected app in Power Apps Studio, and AI code generation tools let makers build one locally through a coauthoring session.
A plan in Power Apps can open a model-driven app in the modern app designer with tables already added, ready to save and publish. The model app-builder skill creates the app and its Dataverse artifacts from a description, after the maker reviews the app spec and approves the plan. Generative pages add tailored experiences from natural language.
The Power Apps vibe experience generates plans, data and code-based apps from natural language in one chat-driven workspace, and Plan mode proposes a plan before building. Draft tables stay in memory until they are published. Microsoft now points app building towards Copilot Cowork and Copilot Studio as the experiences to use going forward.

4. Demonstrate the capabilities of Power Automate

20–25%

Describe Power Automate use cases and capabilities

Cloud flows start from an event (automated), a button (instant) or a schedule (scheduled). Desktop flows use RPA to automate tasks on a workstation, including legacy applications without an API, in attended or unattended mode on physical or virtual machines, and a cloud flow can start them.
Power Automate automates approvals (single response, everyone, custom or sequential), starts flows from Teams messages and posts adaptive cards. It integrates with Outlook, SharePoint lists and libraries, and Forms responses. With AI Builder it extracts data from documents, and the automation toolkit combines Power Automate, Power Apps and Dataverse.

Describe how to build basic flows by using Power Automate

A connector gives access to a service and offers triggers, the events that start a cloud flow, and actions, the steps it performs next. Dynamic content reuses outputs of earlier steps, and conditions and loops control the logic. Connections hold user access, and custom connectors cover APIs without a prebuilt connector.
Copilot in cloud flows creates a flow from a natural language description: it suggests a trigger and actions, sets up connections and lets you refine the result in conversation. In the new designer it also edits existing flows, answers questions about them, creates or fixes expressions and helps resolve errors; its changes should always be reviewed.
Copilot in Power Automate for desktop creates flows or adds actions from a description, inserts a placeholder for recording UI tasks and answers product questions. It also summarizes actions and subflows, generates scripts and Power Fx expressions from natural language and helps repair selectors at run time; everything it generates should be reviewed.

5. Describe features and capabilities of agents in Microsoft Copilot Studio

20–25%

Describe how to build agents in Copilot Studio

Copilot Studio agents serve employees and customers in cases such as IT support, HR questions, contact centers, invoices, procurement and field service. They can answer from organizational knowledge, act in business systems and escalate to people. Some work autonomously, for example when an email arrives, and their value is measured with KPIs and a baseline.
A topic defines a portion of a conversation through nodes on a canvas: messages, questions, conditions, redirects and tools. With generative orchestration the agent picks a topic by its name and description; with classic orchestration, by its trigger phrases. System topics cover essential behaviors and cannot be created or deleted.
Knowledge sources ground agent responses in enterprise data: public websites, documents uploaded to Dataverse, SharePoint, Dataverse tables and enterprise data through connectors. They are added at the agent level or at the topic level with a generative answers node. SharePoint and Dataverse surface only the content each user can access.
Tools let an agent act on external systems: connectors, agent flows, prompts, REST APIs, Model Context Protocol and computer use. Generative orchestration selects them by their description, and they can also be called from a topic. Agent flows are deterministic, while an MCP server offers tools and resources that update dynamically.
An agent must be published at least once before channels are added, and each new publish updates it on all connected channels. Channels include Teams and Microsoft Copilot, websites, mobile apps, Facebook and Azure Bot Service, which requires development. The demo website is for internal testing, not production, and policies can block channels.

Describe how to manage agents in Copilot Studio

Microsoft Agent 365 lets an organization observe, govern and secure its agents, whether built in Copilot Studio, Azure Foundry or third-party platforms. Its registry centralizes the inventory in the Microsoft 365 admin center, and it relies on Microsoft Entra for identity, Purview for data and Defender for threats.
The Monitor page in Copilot Studio summarizes an agent's usage and adoption: active users, session outcomes, reactions, satisfaction, themes, and tool and knowledge use. Agents with event triggers are analyzed through run outcomes and trigger use, and the Savings calculator estimates return on investment.
Agent evaluation in Copilot Studio provides automated, repeatable testing through sets of test cases, either single response or conversation. Each test method measures something different, such as general quality, meaning, keywords, tools used or your own criteria. Rerunning the same test set lets you compare results after a change.

Practice this certification