Cheat sheet
PL-900 Cheat Sheet
Business Value of Power Platform
- 5-10% of exam
- Business Value Drivers Business Value Themes Power Apps vs Automate Citizen Developer
Manage the Environment
- 20-25% of exam
- Dataverse Basics Forms & Views Environment Types Security Roles DLP & ALM
Power Apps Capabilities
- 20-25% of exam
- App Types Canvas Building Blocks Model-Driven Components AI App Building
Power Automate Capabilities
- 20-25% of exam
- Flow Types Flow Building Blocks Connector Tiers M365 Automation
Copilot Studio Agents
- 20-25% of exam
- Agent Core Blocks System Topics Knowledge Sources Tools & Management
Quick Facts
- Exam PL-900
- Credential Power Platform Fundamentals
- Time 45 min
- Questions 40-60
- Pass 700/1000
- Cost $99 USD
- Level Fundamentals, no prereqs
- Blueprint Jul 24 2026
Power Apps vs Power Automate BU
Power Apps
- Builds custom applications
- Replaces paper and spreadsheets
- The user-facing interface
Power Automate
- Automates business processes
- Reduces manual, repetitive errors
- The behind-the-scenes workflow
- Interface vs process automation
Business Value Drivers BU
- Power Apps Fast custom app delivery
- Power Automate Automates repetitive processes
- Dataverse Secure shared data layer
- Connectors No-code service integration
- Power Pages External-facing low-code websites
- Generative AI features Lower platform skill barrier
- Copilot Studio Agents extend support capacity
- Citizen developer Non-IT business app builder
Business Value Themes BU
- Speed to solution Days not months
- Lower dev cost Fewer specialized developer hours
- Data silos Scattered disconnected team data
- Fusion teams Citizen plus pro developers
Environment Types EN
- Default, Production, Sandbox, Trial, Developer
- Default: auto-created, avoid it Production: live, tested apps Sandbox: resettable dev space Trial: time-limited evaluation Developer: free personal space
Business Rule vs Cloud Flow EN
Business rule
- No-code, form-scoped logic
- Require or hide fields
- Runs inside one form
Cloud flow
- Broader multi-step process
- Triggered by Dataverse events
- Notifies or updates records
- Form logic vs full process
Which Environment Type EN
- Live users, business-critical work → Production ( Tested, approved solutions )
- Team testing, needs a reset → Sandbox ( Copy from Production )
- Short evaluation of the platform → Trial ( Expires if not converted )
- Solo maker's personal space → Developer ( Free, one Dataverse database )
- Auto-created at tenant setup → Default ( Avoid building here )
Dataverse Basics EN
- Table Rows and columns
- Column One data field
- Primary column Required unique row identifier
- Standard table Prebuilt Account, Contact tables
- Custom table Org-specific data model
- Lookup column Implements one-to-many relationship
- One-to-many relationship One parent, many children
- Many-to-many relationship Rows relate both directions
Security Access Levels EN
- Access: None, User, BU, Parent-Child, Org
- None: no access at all User: only owned records BU: same business unit Parent-Child: unit plus children Org: every record, period
Unmanaged vs Managed Solution EN
Unmanaged solution
- Fully editable components
- Used during active development
- Lives in Dev environment
Managed solution
- Locked and read-only
- Deployed to Test, Production
- Cleanly upgraded or removed
- Building vs deploying a solution
Forms & Views EN
- Main form Full record view edit
- Quick Create form Fast new record entry
- Quick View form Read-only related record snapshot
- Card form Compact mobile-friendly layout
- System view Ships with the table
- Personal view User's own saved list
- Public view Shared with other users
- Default view Auto-shown record list
Business Logic & AI Schema EN
- Business rule No-code if-then form logic
- Power Fx formula column Real-time calculated value
- Workflow or cloud flow Automates Dataverse-triggered processes
- Plug-in Pro-developer server-side code
- Copilot schema generation Natural-language table creation
- Schema review Confirm accuracy before publishing
Environment Types EN
- Default Auto-created tenant environment
- Production Live business-critical apps
- Sandbox Resettable dev and test
- Trial Time-limited evaluation space
- Developer Free personal maker environment
Security Roles & Scope EN
- Business unit Models org division scope
- Security role Bundle of table privileges
- None access level No access at all
- User access level Only records user owns
- Business Unit access level Records in same unit
- Parent Child access level Unit plus child units
- Organization access level Every record, any owner
Teams & Admin Roles EN
- Owner team Can own and share records
- Entra ID security team Membership synced from Entra
- Access team Shares records, owns none
- Field-level security Protects one sensitive column
- Environment Admin Controls a single environment
- Power Platform Administrator Controls the entire tenant
DLP & ALM EN
- DLP policy Blocks risky connector combos
- Business connector group Approved for org data
- Non-Business connector group Not approved for mixing
- Blocked connector Use prevented entirely
- Tenant-wide DLP Default across all environments
- Environment-scoped DLP Overrides the tenant default
- Unmanaged solution Editable, active development state
- Managed solution Locked, ready for deployment
- Power Platform pipelines Automates Dev, Test, Prod
- CoE Starter Kit Free tenant governance toolkit
App Type Picker PO
- Canvas is UI, Model-Driven is Data, Code is Pro
- Canvas: blank, custom design Model-driven: generated from data Code app: pro developer React
Canvas vs Model-Driven Apps PO
Canvas app
- Blank design surface
- 1,000+ connectors available
- Pixel-level layout control
Model-driven app
- Generated from Dataverse
- Auto security and rules
- Responsive by default
- Custom UI vs structured data
Which Power Apps Type PO
- Need a custom pixel UI → Canvas app ( Blank design surface )
- Need many unrelated data sources → Canvas app ( 1,000+ connectors available )
- Structured, data-first business process → Model-driven app ( Generated from data model )
- Need automatic security enforcement → Model-driven app ( Inherits Dataverse roles, rules )
- Pro developer wants standard code → Code app ( Built with React )
- Only a natural-language app idea → Vibe experience ( AI generates first draft )
- Need the data model planned first → Plan Designer ( AI proposes tables first )
App Types PO
- Canvas app Built from blank canvas
- Model-driven app Generated from data model
- Code app Pro-developer React application
- Plan Designer AI plans data structure
Canvas App Building Blocks PO
- Screen One page users navigate
- Control Button, label, gallery, input
- Data source Connected external record source
- Power Fx Excel-like low-code formulas
- Gallery Repeats layout per record
- 1,000+ connectors Very broad data reach
Model-Driven Components PO
- Site map Defines app navigation structure
- Form Layout for one record
- View Filtered, sorted record list
- Business process flow Guides users step-by-step
- Dashboard Visual chart data summary
AI-Assisted App Building PO
- Vibe experience Conversational, iterative app refinement
- Plan Designer step Structured plan before building
- AI app generation Natural language becomes working app
- Human review step Verify data, security, logic
Flow Trigger Types PO
- Automated is Event, Instant is Button, Scheduled is Timer
- Automated: fires on an event Instant: user clicks a button Scheduled: runs on a timer
Cloud Flow vs Desktop Flow PO
Cloud flow
- Runs in Microsoft cloud
- Needs a connector trigger
- Modern, API-based services
Desktop flow
- RPA on the desktop
- No API or connector
- Automates legacy applications
- Connector exists vs it doesn't
Which Flow Type PO
- Event in a connected service → Automated flow ( New email or item )
- User must trigger it manually → Instant flow ( Mobile or canvas button )
- Runs on a fixed schedule → Scheduled flow ( Recurrence trigger )
- Legacy app with no API → Desktop flow ( RPA automation )
- Need Approve Reject in email → Send email with options ( Outlook connector action )
- Extract data from an invoice → AI Builder model ( Document processing )
Flow Types PO
- Automated flow Starts from a connector event
- Instant flow Starts from a manual click
- Scheduled flow Starts on a recurring timer
- Desktop flow RPA for legacy apps
Standard vs Premium Connectors PO
Standard connector
- Included in most licenses
- Outlook, SharePoint, Teams
- No extra cost
Premium connector
- Requires extra licensing
- Dataverse, SQL Server, Salesforce
- Higher-tier plan needed
- Bundled vs paid add-on
Flow Building Blocks PO
- Trigger Exactly one per flow
- Action Step performed after trigger
- Condition If, then branching logic
- Apply to each Loops over a collection
- Parallel branch Runs steps simultaneously
- Expression Transforms data between steps
Connector Tiers PO
- Connector Prebuilt bridge to a service
- Standard connector Included with most licenses
- Premium connector Requires extra licensing
- Custom connector Built for an internal API
M365 & Document Automation PO
- Approvals connector Sequential or parallel routing
- Teams action Posts channel notifications
- Send email with options Embeds Approve or Reject
- SharePoint trigger Item or file created
- Forms trigger New survey response submitted
- AI Builder extraction Pulls fields from documents
Copilot Studio Building Blocks CO
- Topics, Knowledge, Tools, then Channels
- Topics: scripted conversation paths Knowledge: grounded generative answers Tools: let the agent act Channels: where agents are published
Topic vs Knowledge Source CO
Topic
- Maker-authored, scripted dialog
- Started by trigger phrases
- Predictable exact wording
Knowledge source
- AI-generated at run time
- Grounded in real content
- Handles unscripted questions
- Scripted path vs generative answer
Which Copilot Studio Tool CO
- Standardized external tool access → MCP server ( Open protocol )
- Callable steps inside a conversation → Agent flow ( Deterministic mid-chat step )
- Question matches no topic → Knowledge source ( Generative grounded answer )
- Need an exact scripted path → Topic ( Trigger phrases required )
Agent Core Building Blocks CO
- Topic Structured, maker-authored conversation path
- Trigger phrase Sample utterance that starts topic
- Knowledge source Grounds generative agent answers
- Tool Lets an agent take action
- Channel Where a published agent lives
System Topics CO
- Greeting Agent's opening message
- Escalate Hands off to a human
- Fallback No topic or knowledge matched
- Multiple Topics Matched Asks user to clarify
Knowledge Sources CO
- SharePoint Internal sites and documents
- Public website Company help-center pages
- Uploaded file PDF, Word, or spreadsheet
- Dataverse table Structured business data source
Tools & Agent Management CO
- MCP server Standardized external tool access
- Agent flow Callable steps inside an agent
- Microsoft Agent 365 Tenant-wide agent governance
- Monitoring & analytics Session, resolution, escalation rates
- Agent evaluations Pre-publish regression testing
Common Traps
Canvas vs Model-Driven
- Canvas starts from blank design ≠ Model-driven starts from data
Cloud Flow vs Desktop Flow
- Cloud flow needs a connector ≠ Desktop flow automates legacy apps
Topic vs Knowledge Source
- Topic dialog is maker-scripted ≠ Knowledge answers are AI-generated
Business Rule vs Cloud Flow
- Rule enforces logic on forms ≠ Flow automates a broader process
Unmanaged vs Managed Solution
- Unmanaged stays editable in dev ≠ Managed is locked for deployment
AI Builder vs Copilot Studio
- Builder embeds models in apps ≠ Studio builds conversational agents
MCP Server vs Agent Flow
- MCP standardizes external tool access ≠ Agent flow runs steps in-agent
Environment vs Microsoft Agent 365
- Environment secures one data container ≠ Agent 365 governs agents tenant-wide
Last Minute
- 1. Domains: Value 5-10%, rest 20-25%
- 2. Canvas = blank; Model-driven = data
- 3. Cloud flow = connector; Desktop = RPA
- 4. Automated=event, Instant=click, Scheduled=timer
- 5. DLP blocks Business + Non-Business mix
- 6. Managed = locked; Unmanaged = editable
- 7. Lookup column = one-to-many relationship
- 8. Topics = scripted; Knowledge = generative
- 9. MCP = standard tool access
- 10. Agent 365 = tenant-wide governance
- 11. Pass = 700/1000 scaled score
- 12. PL-900 needs zero coding skill
- 13. Admin: one env; Administrator: whole tenant
Microsoft Power Platform Fundamentals (PL-900)
- Practice Questions 223 questions Study Guide 20 topics Cheat Sheet 119 facts Flashcards 50 cards Videos 1 video Blogs 1 blog Free CE Podcast