AI-R in 99 Steps

Product audit, gap analysis & implementation plan - April 12, 2026

"Ordered by pain level, not by our architecture. What keeps restaurant owners up at night comes first."

AI-R Surface Architecture
/pwa — Restaurant View
Staff-facing. iPad station + phone server mode. Order entry, KDS, timeclock, floor plan, service intel. PIN-gated by role.
/home + /settings — Owner/GM View
Dashboard, reports, scheduling, menu management, integrations, food cost, Maestro AI. Clerk auth (owner/GM only).
/site/[slug] — Customer Restaurant Website
Each AI-R restaurant gets a website on their own custom domain. Menu, hours, photos, online ordering. Public, no auth.
/ + /beta + /demo — Public / Marketing
Landing page, beta signup, demo flow (upload menu → see AI-R in action). Internet-facing, no auth required.
All surfaces read/write to the same Neon Postgres DB via 132+ API routes. One source of truth.
Built & Tested
Built, Needs Testing
Gap - Must Build
Future / Post-Launch

Scoreboard

0
Built & Tested
0
Needs Testing
0
Gaps
0
Future
🔥 Nightmare #1: The Menu Steps 1-14
Every restaurant owner's first question: "How do I get my menu in there?" If this is painful, they're gone.

PHILOSOPHY: The goal is to get the menu 80-90% right on first upload. Owner fixes the small stuff later in the editor. AI gets to 100% AFTER the owner supplies more data (recipes, photos, corrections). Do NOT let the user obsess over perfection during onboarding. Fast to operational > pixel-perfect menu.
1
Upload Menu - Photo or PDF
Owner takes a photo of their paper menu or uploads a PDF. AI extracts everything.
🏆 BIG WIN: No spreadsheet. No manual entry. AI gets 80-90% right in seconds. Toast/Square require manual item-by-item entry.
Done
2
AI Menu Parsing (OCR + GPT)
GPT-4o extracts categories, item names, descriptions, prices, modifiers from the uploaded image/PDF.
Done
3
Menu Review & Correction UI
Owner reviews AI-parsed menu. Fix names, adjust prices, reassign categories. Inline editing.
Needs Test
4
Category Management
Create, rename, reorder, delete categories. Drag items between categories.
Done
5
Item Details - Price, Description, Image
Edit any item. Price, description (AI-generated option), photo upload or AI-generated image.
Done
6
Modifier Groups
Create modifier groups (Size, Temperature, Add-ons). Assign to items. Required vs optional. Price adjustments.
Done
7
Addon Groups
Separate from modifiers - extras, sides, drinks. "Would you like to add...?"
Done
8
Menu Health Score
AI scores menu completeness: missing descriptions, pricing gaps, items without images, category balance.
⚠️ HIDE IN DEMO. Overcomplicates onboarding. Risk: owner obsesses over small details → tries AI fix → AI can't nail it without more data → "AI is no good." Wrong usage at this phase.
Cut from Demo
9
AI Menu Descriptions
Owner writes a basic description → AI enhances it. Or: upload a photo of the dish → AI vision analysis → generates description from what it sees. AI as enhancer >> AI as creator.
⚠️ SIREN SONG: Pure AI-generated descriptions from just a name will be generic and wrong. Owner must supply core data first, THEN AI enhances. Fallback: AI creates from scratch is fine, but set expectations. Best path: photo of item → vision → description.
Done
10
AI Food Photography
AI generates professional food photos from item names. Image wizard with style options.
🏆 Professional photos without hiring a photographer.
Done
11
86 System (Mark Unavailable)
One tap: item is 86'd. Propagates to POS, website, online ordering, delivery apps.
⚠️ Table stakes - every POS does this. Not a differentiator. Must work flawlessly but don't oversell. The real win is propagation to website + 3rd party apps (most POS only 86 on the POS itself).
Done
12
Daily Specials + Pre-Shift Briefing
Manager/chef sets up daily intel: specials (with countdown timers + special pricing), 86 list, push items (with reasons), wine/drink pairings, house notes. All visible to staff on the Service Intel tab during service.
🏆 Full pre-shift briefing system. Replaces PreShifter ($99/mo). Specials, push items, pairings, house notes - all in one place.
Done
13
Channel Pricing
Each location has 1 menu with 1 tab per channel (dine-in, online, DoorDash, UberEats, etc). Different prices per channel to protect margins on delivery. Time-of-day routing for phone answering, happy hours, discount specials.
i️ KEY ARCHITECTURE: 1 location = 1 menu = N channel tabs. Primarily a 3rd party delivery concern but applies across all channels.
Done
14
Kitchen Translations
Auto-translate menu items + modifiers into kitchen staff's language on tickets.
Done
🔥 Nightmare #2: Table Map & Floor Plan Steps 15-22
"Where are my tables? What's open? Who's sitting where?" The spatial reality of a restaurant.
15
Upload Floor Plan Image
Owner takes a photo or screenshot of their existing floor plan and uploads to AI-R. Image becomes the base layer for table placement. Fastest path to operational.
🏆 Use what you already have. No drawing required. Photo → upload → place tables on top.
Done
16
Place Tables on Map
Drag tables onto floor plan. Set table numbers, capacity, shape.
Done
17
Table Status - Live View
Real-time status: open (gray), seated (blue), ordered (yellow), paying (green). Visual at a glance.
Done
18
Tap Table → Start Order
Tap any table on the map to open it, seat guests, start an order.
Done
19
Section Assignments
Assign tables to server sections. Visual section boundaries on the map.
Done
20
Floor Summary - Covers, Turn Time
How many covers? Average turn time per table? Which sections are hot?
Done
21
Multiple Floor Plans (Tabs)
Patio vs indoor. Brunch layout vs dinner layout. Tabs on the page to switch between floor plans.
Needs Test
🔥 Nightmare #3: Printers & KDS Steps 23-32
"The printer isn't working" - the five words that stop a restaurant dead. KDS is the heart of the kitchen. This has to be bulletproof.
23
Printer Discovery (Network Scan)
Auto-discover printers on the restaurant's network. No manual IP entry.
Done
24
Add & Name Printers
Name each printer: "Kitchen," "Bar," "Expeditor." Assign to station.
Done
25
Test Print
One-tap test print from settings. Confirms connection before service.
Done
26
Category-Based Print Routing
Drinks → bar printer. Food → kitchen printer. Desserts → pastry station. Automatic.
Done
27
Receipt Printing
Customer receipt to designated receipt printer. Smart receipt with QR code.
Done
28
Raspberry Pi Print Server
Pi acts as print server bridge. Cloud POS → Pi → local printers. Works with any ESC/POS printer.
🏆 Use any cheap thermal printer. No proprietary hardware lock-in.
Done
29
Print Queue & Retry
If printer is offline, orders queue. Auto-retry when connection restores. Nothing gets lost.
Done
30
Printer Status Monitoring
Dashboard shows printer status: online, offline, paper low. Alert if printer drops mid-service.
Needs Test
31
Kitchen Display System (KDS)
Kitchen sees orders on screen. Bump when done. Priority bumps. Flags for issues. Void requests with manager approval. The heart of the kitchen - everything flows through here.
Done
32
Smart Receipts - Revenue Engine on Paper
Every receipt is tailored by order type. Dine-in: rating → review gate (happy → Google, unhappy → private feedback) + loyalty email capture + referral link. Delivery/To-Go: QR code → "Reorder same meal at 20% discount" → captures phone → auto-scheduled follow-up SMS. Every receipt is a marketing touch.
🏆 MAJOR DIFFERENTIATOR: Receipt becomes a revenue recovery tool. Review gating replaces Birdeye/Podium ($200-300/mo). Reorder discount drives repeat business. Tailored per order type - nobody else does this.
🎯 STRATEGY: Use DoorDash/UberEats as customer ACQUISITION tools. Once a delivery customer gets the smart receipt, convert them to direct ordering. NOT because "we save 30% on commissions" (avoid this claim always) - but because direct = better prices for the customer + no DD/UE fees. Customers order MORE when they pay less. The receipt is the conversion mechanism: DD/UE customer → reorder QR → direct order → lower price → bigger order → repeat.
Done
⚡ Critical #1: Credit Card Processing Steps 31-38
If you can't take money, nothing else matters.

CURRENT: Square (most restaurants) + Stripe (online ordering) + Shopify (OB built early, near-zero restaurant adoption but potential partnership play).
GAP: Need to support any processor eventually. May need a gateway/abstraction layer. Not critical for launch but on the roadmap.
31
Connect Square Account
Owner connects their Square account from settings. Location ID + terminal setup. Tested successfully at Red Bar Sushi.
⚠️ Currently manual Location ID entry. Works but needs OAuth for true self-service at scale.
Done
32
Pair Square Terminal
Generate pairing code in AI-R → enter on Square Terminal → paired. Supports multiple terminals.
Done
33
Process Payment - Tap/Insert/Swipe
Send payment to terminal. Customer taps/inserts card. Confirmation returns to POS.
Done
34
Tip on Terminal
Tip prompt on the Square terminal. Pre-set percentages or custom amount.
Done
35
Tip Adjustment (Post-Payment)
Capture tip from signed receipt. Adjust within settlement window.
Done
36
Split Payments
Split check by item, by seat, by equal share, or by custom amount.
Needs Test
37
Stripe Terminal (Alternative)
Full Stripe Terminal integration. Primarily for online ordering payments. Rarely used for in-restaurant swiped transactions.
Needs Test
37b
Shopify Integration
OB built early integration. Near-zero restaurant adoption currently, but potential partnership play. Could be promoted to get Shopify on our team.
Needs Test
37c
Universal Payment Gateway
Abstraction layer to support any credit card processor. Restaurants should bring their own processor, not be locked into ours.
⚠️ Not critical for launch. But needed to scale beyond Square/Stripe restaurants. Research gateway services that bridge to any processor.
Future
38
End-of-Day Settlement Report
Total payments, tips, voids, refunds. Reconcile with processor.
Needs Test
⚡ Critical #2: Website & Online Ordering Steps 39-46
Free website + zero-commission online ordering = instant value the owner can see.

PHILOSOPHY: Emulate Owner.com's early playbook: "Don't worry about what the site looks like - we're going to get you great SEO." Speed to live > pixel-perfect design. Site MUST be on the restaurant's existing domain. Orders must work on day one. AI shines here: DNS provisioning, SSL, menu-to-site, SEO - all automated.
39
Auto-Generate Restaurant Website
From menu data, AI builds a complete mobile-first website. Menu, hours, location, photos. Done in demo - proven flow.
🏆 BIG WIN: Free. Replaces Owner.com / Popmenu ($200-500/mo). Don't overthink design - SEO and ordering are what matter.
Done
40
Connect Restaurant's Existing Domain
Restaurant MUST use their existing domain - not a subdomain of ours. AI assists with DNS setup, verification, SSL provisioning. Owner gives us domain, we handle the rest.
⚠️ Critical for SEO - existing domain has authority. New subdomain has zero. This is non-negotiable.
Needs Test
41
Online Ordering - Menu Browse & Cart
Customer browses menu on website. Categories, items, descriptions, photos, modifiers. Add to cart, special instructions. Mobile-optimized.
⚠️ GAP: Demo website generates but can't take orders yet. FIX: Reuse the /pwa order entry flow (MenuGrid + OrderPanel) - it's already polished for servers, just simplify for customers. Same code, customer-facing skin.
Gap
42
Online Ordering - Checkout & Payment
Checkout with Stripe for online payments. No commission. Customer pays, order confirmed.
Gap
43
Online Order → Kitchen
Online orders flow directly to KDS/kitchen printer. Same pipeline as dine-in orders - no separate system.
🏆 Zero commission on your own orders. DoorDash charges 15-30%. Same kitchen flow regardless of order source.
Needs Test
44
Order Confirmation & Tracking
Customer gets confirmation. Estimated ready time. SMS updates.
Needs Test
45
Google Business Profile - Pull Data
Google Places API finds restaurant, pulls place_id, photos (up to 10), ratings. Auto-populates restaurant profile during onboarding.
Done
45b
Restaurant Intel Scan (9 Platforms)
Scans DoorDash, UberEats, GrubHub, Google, Yelp, Instagram, Facebook, TikTok + restaurant's own website. Finds existing presence, menus, ratings, review counts, social handles. All automated during onboarding.
🏆 Instant competitive intelligence. Owner sees their full digital footprint in seconds. Nobody else does a 9-platform scan on signup.
Done
45c
Shopify Theme Integration (Curated Sites)
For owners who want a highly curated website, AI-R integrates with Shopify themes. Traditional design options alongside AI-generated sites.
Needs Test
46
SEO Basics - Auto-Generated
Meta tags, structured data (Restaurant schema), sitemap. AI does it during site generation.
Needs Test
⚡ Critical #3: Third-Party Delivery Integration Steps 47-53
DoorDash, UberEats, GrubHub - love them or hate them, orders need to flow into one system.
47
KitchenHub Integration - Full API Client
OAuth client with auto-refresh. Fetch orders, manage locations, control platform status (toggle DD/UE/GH online/offline). Webhook receiver for real-time order pushes + 15-min cron backup sync. Keys configured, location ID set for Chan Nara.
⚠️ FOR TOMORROW: Verify env vars are in staging/production deployment. Webhook URL must point to production: www.ai-restaurant.net/api/webhooks/kitchenhub. Location ID must match Chan Nara in KitchenHub dashboard.
Done
48
DoorDash Orders → AI-R
Via KitchenHub. DoorDash orders appear in POS, print to kitchen. Platform auto-detected. Same order flow as dine-in.
Test Tomorrow
49
UberEats Orders → AI-R
Via KitchenHub. No separate tablet. Orders flow into one unified queue.
Test Tomorrow
50
GrubHub Orders → AI-R
Via KitchenHub. All delivery channels in one view.
Test Tomorrow
51
Delivery Order Tracking
See all delivery orders in one view. Status, driver, estimated pickup time.
Needs Test
52
Menu Sync to Delivery Platforms
Push menu changes from AI-R → DoorDash/UberEats/GrubHub. One source of truth.
Needs Test
53
Delivery Revenue Reporting
See delivery revenue by platform. Compare commission impact. Know your real margins.
Needs Test
💪 Core POS Operations Steps 54-70
A chronological day at the restaurant. Staff arrives → service → close. This is where staff lives all day.
54
Staff Clock In (PIN Pad Kiosk)
Staff enters PIN on iPad. Warm kiosk UI. Shows today's shifts, message of the day, who's already clocked in. Station select for kitchen staff (grill, fry, sauté, prep, expo, bar). Role determines access level for the rest of the shift.
Done
54b
Staff Management
Add/edit/deactivate staff. Roles (Owner, GM, Manager, Kitchen, Server, Host). PIN assignment. Hourly rate. Permissions by role.
Done
54c
Scheduling - Weekly Grid + Templates
Drag-and-drop weekly schedule. Day timeline view. Save templates for recurring weeks. Apply template with one click. Shift details in popover.
🏆 Replaces 7shifts / Homebase / When I Work ($50-100/mo).
Done
54d
Break Tracking
Start break / end break from timeclock. Break compliance tracking. Manager can edit entries.
Done
54e
Staff Status Board
Real-time view: who's clocked in, who's on break, who's late. Visible on manager tab.
Done
55
Phone Server Mode
Server uses their phone as a POS. Table list, order entry, fire to kitchen. No iPad needed.
🏆 No extra hardware cost per server.
Done
56
iPad Station Mode
Shared iPad at host stand or bar. Swipeable tabs: Today, Orders, Info, Payments, Manager.
Done
57
Order Entry - Menu Grid
Browse categories → tap items → add to order. Fast, visual, touch-optimized.
Done
58
Modifiers on Order Entry
Modifier prompts when adding items. Required modifiers enforced.
Done
59
Fire to Kitchen
Send order to KDS and/or kitchen printers. Coursing support.
Done
60
Kitchen Display System (KDS)
Kitchen sees orders on screen. Bump when done. Priority bumps. Flags for issues.
Done
61
Edit Open Orders
Add items, remove items, change mods on orders that haven't closed.
Done
62
Void Items (Manager PIN)
Void requires manager PIN. Reason code logged. Accountability.
Done
63
Bar Tabs - Full Lifecycle
Open, add items, split, transfer between servers, close out.
Done
64
To-Go / Pickup Orders
Create orders not tied to a table. Name/phone for pickup.
Done
65
Payment → Receipt → Close
Full checkout flow: payment modal → tip → receipt print → table freed.
Done
66
Smart Receipts
Digital receipt via QR. Review prompt, feedback, reorder link. Conversion tracking.
🏆 Every receipt is a marketing touch. Replaces Birdeye/Podium ($200-300/mo).
Done
67
Pending Tips - End-of-Shift Entry
Payments tab shows all pending tips (paper receipt signatures). Each row: time, table, card last 4, authorized amount. Server enters tip from signed paper receipt. Calls Square API to adjust and complete payment. One-by-one or batch capture. Full paper receipt → digital tip workflow.
⚠️ TEST TOMORROW: Verify full flow - payment authorized on terminal → paper receipt signed by customer → server enters tip at end of shift → Square adjusts and completes.
Done
68
Staff Clock Out & Break Tracking
PIN → clock out. Break start/end tracked. Hours calculated.
Done
69
Auto-Return to Today Tab (Idle)
iPad returns to Today dashboard after 60 seconds idle. Prevents stuck screens.
Done
70
Service Intel - Live During Service
Real-time data tab: covers, revenue, avg ticket, top sellers. Visible to staff on iPad.
Done
💪 Back Office & Intelligence Steps 71–85 — Test at Chan Nara this week
Where the owner makes money decisions. The intelligence layer that makes AI-R different.

DATA ARCHITECTURE: All real customer data lives in Neon Postgres (132 DB-backed API routes). 33 localStorage stores exist from prototype/testing phase - these are scaffolding, not production. Critical path (order → kitchen → payment) is fully DB-backed. Cleanup: migrate remaining localStorage stores to DB as features go to production. Goal: staff enters data anywhere (KDS, iPhone, iPad) → DB → instantly available to all systems.
71
Owner Dashboard
Daily summary: revenue, orders, avg ticket, top sellers, labor cost. Real-time.
Done
72
Order History & Search
All past orders. Filter by date, server, table, status. Full detail drill-down.
Done
73
Scheduling - Weekly Grid + Templates
Drag-and-drop schedule builder. Save templates. Recurring schedules.
🏆 Replaces 7shifts / Homebase ($50-100/mo).
Done
74
Timeclock Reports
Hours per staff member. Overtime alerts. Break compliance. Export for payroll.
Done
75
Tip-Out Calculation
Auto-calculate tip pools by section. Fair, transparent, documented.
Done
76
Recipe Costing & Plate Cost
Define recipes → auto-calculate cost per plate → see real margins.
Done
77
AI Recipe Generation
AI creates baseline recipes from menu item names. Instant food cost estimates.
🏆 Zero-to-food-cost in minutes. Nobody else does this.
Done
78
Invoice Scanning (OCR)
Photo supplier invoices → AI extracts items, prices, vendors. No manual entry.
Done
79
Inventory & Low-Stock Alerts
Track current stock. Auto-alerts when items hit reorder point.
Done
80
Food Cost Variance
Theoretical vs actual cost. Find waste, theft, over-portioning.
Done
81
Maestro - AI Consultant
Three lenses: CFO (margins), Ops (efficiency), Marketing (revenue). Ask anything in natural language.
🏆 An AI consultant that never clocks out. $0 extra. This is the moat.
Done
82
QuickBooks Export
Push financial data to QuickBooks. QB becomes compliance tool, not management tool.
🏆 Replaces daily QuickBooks wrestling.
Needs Test
83
Plaid - Bank Account Link
Connect bank account for cash flow visibility. Real-time financial picture.
Needs Test
84
Daily Digest (Auto-Report)
Automatic daily summary: revenue, labor, food cost, highlights. Delivered to owner.
Done
85
Upsell Tracking & Analysis
AI identifies upsell opportunities. Track staff execution. Measure revenue impact.
Done
🚀 Growth & Future Steps 86-99
Features that differentiate and scale. Build after core is bulletproof.
86
Phone AI - Voice Ordering
AI answers the restaurant phone. Takes orders via natural conversation. Sends to kitchen.
🏆 100% of calls answered. No missed revenue. Nobody else has this.
Needs Test
87
Review Management & Gating
Collect feedback on every receipt. Route happy customers to Google. Intercept unhappy ones privately.
Needs Test
88
Social Media Posting
AI generates social posts from specials, events, food photos. Post to Instagram/Facebook.
Needs Test
89
Meta Ads Integration
Run targeted ads from AI-R. Track conversions back to orders.
Needs Test
90
Chargeback Recovery
AI analyzes chargebacks. Auto-generates dispute documentation. Recovers revenue.
Needs Test
91
Pre-Shift Briefings
Auto-generated: specials, 86 list, push items, wine pairings, VIP notes.
🏆 Replaces PreShifter ($99/mo).
Needs Test
92
Reservation System
Accept reservations on website. Manage from POS. Table assignment on arrival.
Needs Test
93
Waitlist Management
Add walk-in guests to waitlist. Estimated wait time. SMS notification when table is ready.
Needs Test
93
Customer Loyalty / Reorder
Smart receipt drives reorders. Track repeat customers. Loyalty rewards.
Future
94
Multi-Location Dashboard
Owners with multiple restaurants see all locations in one view. Compare performance.
Future
95
Ambient Audio Monitoring
Listen to front-of-house audio. AI identifies service issues, noise levels, wait times.
Needs Test
96
Square OAuth (Self-Service Connect)
Owner clicks "Connect Square" → logs into Square → auto-configured. No manual IDs.
Gap
97
Automated Payroll Export
Export timeclock data directly to payroll provider. Eliminate manual entry.
Future
98
Orion Integration - AI Growth Engine
AI-R runs the restaurant. Orion grows it. Marketing, SEO, reputation - the intelligence add-on.
Future
99
API - Open Platform
Public API for third-party integrations. Let the ecosystem build on AI-R.
Future
Last updated: April 12, 2026 · Built by Dave & Neil