Ce contenu n’est pas encore disponible dans votre langue.
Overview
:::info IntelliKDS is an optional add-on to the standard KDS module and requires an active KDS subscription. It cannot be used as a standalone module. :::
IntelliKDS is an AI-driven extension of the standard KDS that replaces estimated preparation times with a real-data forecast model. Instead of reacting to delays after they happen, the system calculates when each ticket needs to enter the kitchen — and dispatches it automatically at the right moment. Service defines a target serving time; IntelliKDS handles everything in between. The model learns continuously after every shift, accounts for factors like staffing levels, table size, and concurrent order density, and becomes more precise over time without any manual reconfiguration.
Upgrading from the standard KDS to IntelliKDS requires no hardware changes and no workflow retraining — the kitchen continues operating on the same screens and confirmation logic.
Core IntelliKDS Features
Feature | Description |
AI-based preparation forecast | Forecasts preparation time per ticket from real operational data — not static estimates — factoring in table size, kitchen staffing, time of day, and concurrent order load |
Automatic kitchen dispatch | Service sets a target serving time; the system calculates the kitchen start and sends the ticket automatically at the right moment without manual intervention |
Over-ticketing protection | Detects approaching kitchen capacity overload and automatically shifts tickets into the next available time slot before a bottleneck occurs |
Simultaneous full-evening scheduling | Target serving times are calculated for all tables across the entire service at once — not reactively table by table |
Continuous self-learning | The model updates after every shift using real outcome data; accuracy improves automatically without manual configuration |
Live Scheduling Dashboard | Real-time view of all active tickets with table, course, target serving time, calculated kitchen start, and current status (Scheduled / Sent / Confirmed) |
Manual override per ticket | Service or the head chef can override the AI suggestion for any individual ticket at any time — e.g. to prioritize a VIP table |
Emergency instant-send button | Bypasses the AI dispatch for a specific ticket with a single tap — for urgent situations without disabling the module |
Shift-level adjustment sensitivity | Per-tenant configuration of how aggressively the system shifts tickets — some operations prefer slightly early over slightly late |
Shift-type recognition | Distinguishes Lunch vs. Dinner, weekday vs. weekend, and high-load evenings — separate patterns per service context |
Seasonal pattern memory | Recognizes and retains seasonal operating patterns (e.g. summer terrace vs. winter indoor) rather than overwriting them with the most recent shift data |
Cold-start fallback | New tenants without shift history fall back to standard KDS dispatch behavior until enough shifts have been logged for the model to activate |
Special event / banquet detection | Distinguishes regular service from group reservations or banquet-style sequences to prevent event data from distorting the standard learning curve |
Confidence score per prediction | Every dispatch decision carries a confidence indicator — lower on atypical days (events, unusual occupancy), higher on stable repeating patterns |
Shift-level decision transparency | Each ticket displacement shows a brief plain-language explanation in the dashboard so staff understand why a ticket was moved |
Manual shift-end feedback | After each shift, staff can rate prediction accuracy (too early / too late / accurate) to accelerate model improvement beyond automatic learning |
How Automatic Dispatch Works
When a course is ordered at the POS, IntelliKDS does not send the ticket to the kitchen immediately. Instead:
Service sets a target serving time for the table and course — either at the time of ordering or as a per-table default for the evening
The model calculates the kitchen start time based on the forecast preparation duration for the specific ticket under the current conditions
The ticket is held in Scheduled state and appears in the Live Scheduling Dashboard
At the calculated start moment, the ticket is dispatched automatically to the relevant kitchen station
Status updates in real time: Scheduled → Sent → Confirmed (upon kitchen confirmation via KDS)
No manual intervention is required between step 1 and step 5 under normal service conditions.
Over-Ticketing Protection
When multiple tickets are scheduled to enter the kitchen at the same time and the model detects that the combined load exceeds capacity, the system automatically distributes the affected tickets across the next available slots:
The shift is calculated based on current kitchen staffing, station load, and historical throughput at that time of day
The displaced ticket is flagged in the dashboard with the reason for the shift (e.g. "Moved: concurrent load on grill station exceeded capacity threshold")
Service sees the updated target time immediately and can override if needed
The number of prevented over-ticketing situations is tracked as a KPI
Forecast Factors
IntelliKDS evaluates the following signals when calculating a preparation forecast:
Factor | Description |
|---|---|
Table size / party size | Larger parties typically require longer plating windows |
Kitchen staffing level | Active station headcount at the time of dispatch is factored into throughput capacity |
Time of day | Peak-hour throughput differs from off-peak throughput even for the same dish |
Concurrent order density | The number of tickets already in the kitchen or scheduled to enter simultaneously |
Historical shift data | Actual confirmed preparation times for each dish per station over prior shifts |
Service context | Lunch vs. Dinner, weekday vs. weekend, high-load vs. standard evening |
Seasonal patterns | Summer / winter or terrace / indoor variations are stored as distinct pattern layers |
Live Scheduling Dashboard
The scheduling dashboard gives service and the head chef a real-time view of the entire evening's dispatch timeline:
One row per active course ticket
Columns: Table — Course — Target Serving Time — Calculated Kitchen Start — Status
Status values: Scheduled (waiting for dispatch) / Sent (dispatched to kitchen) / Confirmed (kitchen has confirmed)
Confidence score indicator per row (high / medium / low)
Displacement reason visible inline when a ticket has been moved
Manual override and emergency instant-send accessible per row without leaving the dashboard
Manual Override and Emergency Dispatch
IntelliKDS is designed to run without daily manual intervention, but two override mechanisms are always available:
Manual override per ticket Service or the head chef can change the kitchen start time or target serving time for any individual ticket at any time. The AI does not reassert itself once a ticket has been manually adjusted for that session.
Emergency instant-send A dedicated button per ticket immediately dispatches the ticket to the kitchen, bypassing all scheduling logic. This is intended for genuinely urgent situations — e.g. a guest who needs to leave early — and does not disable IntelliKDS for any other ticket.
Self-Learning Model
After every shift, IntelliKDS compares its predicted preparation times against the actual confirmed times recorded by the standard KDS. The outcome is used to adjust future forecasts:
Systematic early or late patterns are corrected automatically
Seasonal patterns are stored as distinct model layers — not merged into or overwritten by the current baseline
Special events and group reservations that are flagged as atypical are excluded from standard model updates to preserve accuracy for regular service
Staff can submit optional feedback at end of shift to accelerate corrections for non-obvious edge cases
Cold start behavior: A new tenant without shift history starts with standard KDS dispatch behavior (manual or time-based). Once a minimum number of shifts has been logged, IntelliKDS activates its scheduling logic automatically.
What IntelliKDS Is Not
IntelliKDS does not replace the standard KDS module. It does not change the kitchen monitor view, the confirmation workflow, or the station routing logic. The kitchen team continues to work exactly as before — IntelliKDS operates upstream, determining when tickets are dispatched, not how they are displayed or confirmed.
Standard KDS vs. IntelliKDS:
Standard KDS Add-on | Premium KDS (IntelliKDS) Add-on |
Digital kitchen monitor with ticket view | All Standard KDS features, plus: |
Acoustic and visual alerts | AI-based preparation forecast from real operational data |
Color-coded delay display | Automatic kitchen dispatch at the calculated moment |
Course confirmation for service | Over-ticketing protection with automatic slot distribution |
Station routing and All-Day view | Simultaneous full-evening scheduling across all tables |
Allergen and special-request highlights | Self-learning model per restaurant — no manual reconfiguration |
Recall and partial confirmation | Live Scheduling Dashboard with status, confidence score, and reasons |
Preparation time and delay statistics | Manual override, emergency dispatch, and shift-end feedback |
Dashboard Metrics
IntelliKDS extends the standard KDS dashboard with AI scheduling KPIs:
Forecast accuracy rate — average deviation between predicted and actual serving time, tracked over time
Number of over-ticketing situations prevented (current shift and rolling 30-day total)
Distribution of prediction confidence levels across the current shift (high / medium / low)
Manual override frequency — how often staff adjust AI suggestions per shift
Cold-start status indicator for new tenants (shifts logged vs. minimum required for model activation)
:::danger Pricing
DO NOT QUOTE FROM THIS PAGE Quotable pricing information can be found only at Pricing & Packages :::
:::info Known Issues