Analyzing Diagrams for Shared Functionality («include») and Optional Behavior («extend»)
Once you have an initial Use Case Diagram with actors and use cases connected directly, the next level of maturity is to identify reuse and variability in the model. Two of the most important UML relationships in use case diagrams are:
- «include» — Used when one use case always incorporates the behavior of another use case as a mandatory sub-step. This relationship eliminates duplication by factoring out common, required functionality that multiple use cases share.
- «extend» — Used when one use case optionally adds behavior to another use case at specific, well-defined extension points. This is ideal for conditional, alternative, or exceptional paths that only occur under certain circumstances.
Manually spotting these relationships can be time-consuming and error-prone, especially in larger systems. Visual Paradigm’s AI Use Case Diagram Refinement Tool automates much of this analysis by:
- Scanning the names, actors, and context of your use cases
- Detecting semantic patterns of shared responsibility (for «include»)
- Identifying conditional or variant behavior (for «extend»)
- Proposing directed arrows with proper stereotypes («include» or «extend»)
- Highlighting potential extension points on base use cases
The tool presents suggestions in a review panel or directly on the diagram (with hover explanations), allowing you to accept, reject, or modify each one. Accepted relationships automatically update the diagram with clean, non-crossing arrows and maintain traceability.
Practical Examples
Example 1: GourmetReserve – Mobile Dining Reservation App
AI Analysis & Suggested Relationships:
- «include» relationships (shared mandatory functionality):
- Book a Table «include» Authenticate User → Every booking requires the diner to be logged in or verified.
- Book a Table «include» Process Payment → A deposit or full payment is mandatory to confirm the reservation.
- Pre-order Meal «include» Process Payment → Pre-ordering also requires payment confirmation.
- Cancel Reservation «include» Authenticate User → Only the account holder can cancel.
- «extend» relationships (optional/conditional behavior):
- Apply Discount Coupon «extend» Book a Table → Extension point: “Enter promo code” — only occurs if user provides a valid code.
- Handle Waitlist «extend» Book a Table → Extension point: “No tables available” — adds waitlist registration behavior conditionally.
- Send No-Show Charge Notification «extend» Cancel Reservation → Extension point: “Cancellation within 2 hours of booking” — applies late-cancellation fee logic only in specific cases.
Diagram impact:
- “Authenticate User” and “Process Payment” become central hubs with multiple incoming «include» arrows → dramatically reduces line clutter.
- “Book a Table” oval shows two extension points with dashed «extend» arrows coming in → clearly communicates that discounts and waitlisting are optional enhancements.
Example 2: SecureATM – Next-Generation ATM Network
AI-Identified Relationships:
- «include» (mandatory shared steps):
- Withdraw Cash «include» Authenticate User
- Transfer Funds «include» Authenticate User
- Deposit Check «include» Authenticate User
- Withdraw Cash «include» Validate Transaction Limits
- Transfer Funds «include» Validate Transaction Limits
- «extend» (conditional extras):
- Request Biometric Verification «extend» Withdraw Cash → Extension point: “Amount > $1,000” — adds extra security step only for large withdrawals.
- Print Receipt «extend» Withdraw Cash → Extension point: “User selects print” — optional at the end of most transactions.
- Issue Temporary Card «extend» Report Lost/Stolen Card → Extension point: “Customer requests emergency access” — conditional emergency feature.
Visual benefit:
- Authentication and limit validation are factored out once → reused across many customer use cases.
- High-value transactions clearly show additional security layers via «extend» arrows.
Example 3: CorpLearn – Corporate E-Learning Platform
AI-Proposed Relationships:
- «include»:
- Enroll in Course «include» Authenticate via SSO
- Take Final Assessment «include» Record Learning Progress
- Complete Learning Module «include» Record Learning Progress
- «extend»:
- Request Manager Approval «extend» Enroll in Course → Extension point: “Course cost > $500 or marked as manager-approved only”
- Issue Certificate «extend» Take Final Assessment → Extension point: “Score ≥ 80%” — certificate generation is conditional on passing.
- Send Completion Reminder «extend» Complete Learning Module → Extension point: “Module not finished within assigned deadline” — nudges only when overdue.
Resulting clarity:
- Repeated authentication and progress-tracking steps are modeled once.
- Managers and compliance teams immediately see that approvals and certificates are not automatic—they depend on conditions.
Quick Guidelines for Reviewing AI Suggestions
- «include» → Ask: “Is this behavior always required for the base use case to succeed?”
- «extend» → Ask: “Does this behavior happen only sometimes, at a specific point, and add optional/conditional value?”
- Reject or adjust if the AI misinterprets domain rules (e.g., suggesting «include» for something that’s actually optional).
- After accepting, check that extension points are clearly named on the base use case oval.
By the end of this intelligent relationship identification step, your Use Case Diagram evolves from a simple “who does what” list into a structured model that explicitly expresses reuse, dependencies, and variability. This significantly reduces redundancy, improves maintainability, and prepares the ground for writing precise, non-repetitive flow-of-events descriptions in the next module. The AI handles pattern detection—you bring the business judgment to make the relationships meaningful.