# 3. Career Recommendation Engine

User Type: **Super Administrator**
Source: *Mi Digital Academy - Education CRM Features Document*

---

## 3. Career Recommendation Engine

### 3.1 Recommendation Rules
**What it does:** The Super Admin defines the rules for the career recommendation engine so recommendations are generated consistently.

**Sub-features:**
- Rule creation: the creation of a recommendation rule
- Rule conditions: the conditions that trigger a recommendation
- Rule actions: the recommendation produced by the rule
- Rule priority: the priority order of the rules
- Rule enablement: the rule toggled active or inactive
- Available on web
- Event logging (action performed)
- Audit logging of recommendation rules

**Super Admin User Journey:**
1. As a Super Admin, open the recommendation rules.
2. Create a rule with conditions and actions.
3. Set the rule priority.
4. Activate the rule and verify it is applied.

**Rules & Edge Cases:**
- A rule creation requires at least one condition and one action.
- A rule condition is a learner attribute or aptitude threshold.
- A rule action is the recommendation to produce.
- A rule priority orders the rules when multiple match.
- A rule enablement controls whether the rule is applied.
- A rule with no matching condition produces no recommendation.
- Rule actions are audit-logged with the actor and timestamp.

### 3.2 Strength-Based Matching
**What it does:** The Super Admin configures the strength-based matching so recommendations align with the learner's aptitude strengths.

**Sub-features:**
- Strength mapping: the mapping of strengths to career paths
- Match threshold: the threshold for a strength match
- Match scoring: the scoring of the match strength
- Multi-strength matching: the matching across multiple strengths
- Match preview: the preview of the matches
- Available on web
- Event logging (action performed)
- Audit logging of strength-based matching

**Super Admin User Journey:**
1. As a Super Admin, open the strength-based matching.
2. Map the strengths to career paths.
3. Set the match threshold and scoring.
4. Preview the matches and save.

**Rules & Edge Cases:**
- A strength mapping links a strength to a career path.
- A match threshold is the minimum strength for a match.
- A match scoring ranks the matches by strength.
- A multi-strength matching combines multiple strengths.
- A match preview shows the expected matches.
- A match below the threshold is not recommended.
- Matching changes are audit-logged with the actor and timestamp.

### 3.3 Path Recommendations
**What it does:** The Super Admin manages the path recommendations so learners are guided to the right career paths.

**Sub-features:**
- Recommendation generation: the generation of path recommendations
- Recommendation ranking: the ranking of the recommendations
- Recommendation limit: the number of recommendations shown
- Recommendation context: the context for the recommendation
- Recommendation history: the history of recommendations per learner
- Available on web
- Event logging (action performed)
- Audit logging of path recommendations

**Super Admin User Journey:**
1. As a Super Admin, open the path recommendations.
2. Generate the recommendations for a learner.
3. Review the ranking and limit.
4. Check the recommendation history.

**Rules & Edge Cases:**
- A recommendation generation applies the rules and matching.
- A recommendation ranking orders the recommendations by fit.
- A recommendation limit caps the number shown.
- A recommendation context explains why the path is recommended.
- A recommendation history stores the past recommendations.
- A recommendation with no match shows a fallback message.
- Recommendation actions are audit-logged with the actor and timestamp.

### 3.4 Resource Recommendations
**What it does:** The Super Admin manages the resource recommendations so learners are guided to the study resources that support their path.

**Sub-features:**
- Resource mapping: the mapping of resources to career paths
- Resource relevance: the relevance of the resource to the path
- Resource ranking: the ranking of the resources
- Resource limit: the number of resources shown
- Resource preview: the preview of the resource recommendations
- Available on web
- Event logging (action performed)
- Audit logging of resource recommendations

**Super Admin User Journey:**
1. As a Super Admin, open the resource recommendations.
2. Map the resources to career paths.
3. Set the relevance and ranking.
4. Preview the resource recommendations and save.

**Rules & Edge Cases:**
- A resource mapping links a resource to a career path.
- A resource relevance scores the fit of the resource.
- A resource ranking orders the resources by relevance.
- A resource limit caps the number shown.
- A resource preview shows the expected recommendations.
- A resource with no mapping is not recommended.
- Resource actions are audit-logged with the actor and timestamp.

### 3.5 Recommendation Delivery
**What it does:** The Super Admin manages the delivery of the recommendations so learners receive their career guidance.

**Sub-features:**
- Delivery channel: the channel for the recommendation (in-app, email)
- Delivery timing: the timing of the recommendation delivery
- Delivery frequency: the frequency of the recommendations
- Delivery tracking: the tracking of the delivered recommendations
- Delivery status: the status of the delivery
- Available on web
- Event logging (action performed)
- Audit logging of recommendation delivery

**Super Admin User Journey:**
1. As a Super Admin, open the recommendation delivery.
2. Set the delivery channel, timing, and frequency.
3. Deliver the recommendations.
4. Monitor the delivery tracking and status.

**Rules & Edge Cases:**
- A delivery channel is the in-app or email channel.
- A delivery timing sets when the recommendation is sent.
- A delivery frequency sets how often recommendations are sent.
- A delivery tracking records the delivered recommendations.
- A delivery status reflects the state of the delivery.
- A delivery to a learner with no recommendation is skipped.
- Delivery actions are audit-logged with the actor and timestamp.