Skip to content

Emily Knowledge Requirements ​

What Emily needs to know to effectively assist merchants.

Core Principles ​

1. A Child Should Be Able to Set Up ​

Emily's guidance should be so clear that anyone, regardless of technical expertise, can:

  • Set up a new location from scratch
  • Configure online ordering
  • Create menus and items
  • Set up payments
  • Configure delivery zones
  • Manage capacity limits

2. Deep Understanding, Not Just Steps ​

Emily must understand:

  • What each field does
  • Why it exists (business purpose)
  • How it works (technical implementation)
  • Where it affects other systems
  • When to use different values
  • Who it impacts (staff, customers)

3. Cross-System Awareness ​

Every backoffice setting can affect:

  • Online Ordering (customer web app)
  • Kiosk (self-service terminals)
  • KDS (kitchen displays)
  • Receipts
  • Reports
  • Third-party integrations

Emily must explain these connections.


Knowledge Categories ​

1. Feature Knowledge ​

For each feature, Emily must know:

Feature: [Name]
├── Purpose: Why does this feature exist?
├── Location: Where in backoffice?
├── Fields: Every configurable option
│   ├── Field name and label
│   ├── Data type and validation
│   ├── Default value
│   ├── Dependencies
│   ├── Business logic
│   └── Customer impact
├── Relations: Connected features
├── Examples: Real-world configurations
└── Troubleshooting: Common issues

2. Field Knowledge ​

For each field, Emily must know:

Field: [Name]
├── What: Description of the field
├── Type: Data type (text, number, toggle, etc.)
├── Validation: Rules and constraints
├── Default: Default value
├── Depends On: Parent fields/conditions
├── Business Logic: How it's processed
├── Customer Impact:
│   ├── Online Ordering effect
│   ├── Kiosk effect
│   └── KDS effect
├── Examples: Sample values and use cases
└── Troubleshooting: Field-specific issues

3. Workflow Knowledge ​

Emily must guide through complete workflows:

Workflow: [Name]
├── Goal: What we're trying to achieve
├── Prerequisites: What must be done first
├── Steps: Ordered list of actions
│   ├── Step 1: Navigate to X
│   ├── Step 2: Configure Y
│   └── Step 3: Save and verify
├── Verification: How to confirm success
└── Common Mistakes: What to avoid

4. Troubleshooting Knowledge ​

For each issue, Emily must know:

Issue: [Description]
├── Symptoms: What the user sees
├── Causes: Possible reasons
├── Diagnosis: How to identify the cause
├── Solutions: Step-by-step fixes
└── Prevention: How to avoid in future

Knowledge Depth Levels ​

Level 1: Basic (What) ​

  • Field exists
  • Basic description
  • Where to find it

Level 2: Intermediate (How) ​

  • How to configure
  • Validation rules
  • Default values
  • Dependencies

Level 3: Advanced (Why) ​

  • Business logic
  • Technical implementation
  • Edge cases
  • Performance implications

Level 4: Expert (Impact) ​

  • Cross-system effects
  • Integration impacts
  • Optimization strategies
  • Advanced troubleshooting

Emily should operate at Level 3-4 for all features.


Required Knowledge by Feature ​

High Priority (Must Have Deep Knowledge) ​

  1. Order Capacity Management ✅

    • All 11 fields documented
    • Business logic explained
    • Customer impact detailed
    • Troubleshooting complete
  2. Online Ordering Setup 🔄

    • ~25 fields to document
    • Time calculation logic
    • Scheduling options
    • Payment flow
  3. Delivery Zones

    • Zone configuration
    • Postal code matching
    • Fee calculation
    • Minimum orders per zone
  4. Menu Items

    • Item creation
    • Pricing
    • Availability
    • Allergens
    • Images
  5. Modifiers

    • Group configuration
    • Required vs optional
    • Min/max selections
    • Price adjustments
  6. Locations

    • Profile setup
    • Address configuration
    • Timezone importance
    • Multi-location management
  7. Business Hours

    • Schedule configuration
    • Multiple time ranges
    • Holiday handling
    • Impact on ordering
  8. Payments

    • Stripe setup
    • Payment methods
    • Refund handling
    • Webhook configuration
  9. Loyalty Program

    • Points configuration
    • Earning rules
    • Redemption rules
    • Customer balance
  10. Offers & Promotions

    • Offer types
    • Conditions
    • Usage limits
    • Promo codes

Medium Priority ​

  1. Team Management
  2. Devices (Kiosk, KDS)
  3. Inventory
  4. Categories
  5. Menus
  6. Branding
  7. Receipts
  8. Tax Rates

Lower Priority ​

  1. Translations
  2. Languages
  3. Compliance
  4. Billing
  5. Activity Log
  6. Reports
  7. Integrations (Deliveroo, Uber Eats, etc.)

Each knowledge entry should be structured for optimal search:

jsonc
{
  "id": "unique-identifier",
  "type": "feature|field|workflow|troubleshooting",
  "name": "Human-readable name",
  "name_nl": "Dutch name",
  "name_fr": "French name",
  "name_de": "German name",
  "content": "Detailed content in English",
  "content_nl": "Dutch content",
  "content_fr": "French content",
  "content_de": "German content",
  "category": "features",            // the SOURCE FOLDER name (features, guides, onboarding, ...) — set by extract.js, not a fixed enum
  "route": "/backoffice/path",
  "tags": ["searchable", "keywords"],
  "priority": 90,
  "embedding": [/* vector */]
}

Quality Checklist ​

Before marking a feature as documented:

  • [ ] All fields listed with complete properties
  • [ ] Business logic explained clearly
  • [ ] Customer impact documented for OO and Kiosk
  • [ ] Dependencies mapped
  • [ ] Examples provided
  • [ ] Troubleshooting guides written
  • [ ] Cross-feature relations documented
  • [ ] Content ready for translation

Maintenance ​

When to Update Knowledge ​

  1. New Feature: Document before release
  2. Field Change: Update immediately
  3. Bug Fix: Add to troubleshooting
  4. Customer Feedback: Improve explanations
  5. New Integration: Document connections

Update Process ​

  1. Update markdown documentation
  2. Push to testing / production; the KB workflows extract it and ingest it into the backend KB
  3. Test Emily's responses

There is no translation step — the pipeline does not generate translations.