Full-service restaurant: POS setup, KDS & integrations

1. Sector overview

Full-service restaurants — dine-in with table service, from casual bistros to fine dining — run on coordination between the floor and the kitchen. What makes this sector different:

• Courses and timing matter. Appetizers, mains, desserts, and fired/hold commands mean orders have routing logic, not just item lists.
• Tipping is part of checkout. Pre-auth, tip adjust, and tip pooling make card handling more complex than retail.
• The kitchen is a real-time system. Kitchen display screens (KDS) or kitchen printers decide whether the line cooks in the right order.
• Reservations and walk-ins coexist. Floor plans, table turns, and waitlists are operational tools, not nice-to-haves.
• Margins are thin. A published 2025 industry figure put full-service margins around 2.8% — every fee line item matters more here than in most sectors.

2. In-store hardware & software

3. Flagship sample setup

Scenario: an 80-seat full-service restaurant — dining room, bar, expo window, and a hot line.

Connection map (model-agnostic)

Hardware list

Layout notes

• KDS screens mount where the line can see them without turning — expo screen faces the pass, kitchen screen faces the hot line.
• Kitchen printers need heat-rated paper and a spot away from direct steam; mount high enough that tickets don't get splashed.
• Run ethernet to the kitchen before opening — retrofitting cable through a working kitchen is miserable.
• Handheld docks live at the host stand on chargers; a dead handheld at 7pm is a staffing problem, not a tech problem.
• Keep the switch and router in a locked office; label every cable at both ends.

Daily workflow

1. Pre-service: KDS screens on and showing the right dayparts, printers loaded, handhelds charged, one test fire to the kitchen.
2. Service: orders fired by course; expo bumps completed tickets; handhelds take payment at the table with tip adjust.
3. If the KDS dies: switch routing to the kitchen printer (pre-configured fallback) — don't re-enter tickets by hand mid-rush.
4. Close: settle card batches before the processor cutoff, reconcile tips, run the end-of-day report, charge everything.

4. Setup guide

Restaurants fail POS go-lives on the menu, not the hardware. Do these in order.

1. Build the menu with modifiers first. Every item, every modifier (add chicken, no onion, dressing on side), every forced question — the menu is the database everything else hangs off.
2. Set up taxes. Dine-in vs. takeout can tax differently depending on province and product — confirm with your accountant.
3. Draw the floor and table map. Tables, sections, bar seats, patio — servers can't run a floor the system doesn't understand.
4. Configure printer and KDS routing. Which items fire to the hot line, which to the bar, which to expo. Test every category, not just the common ones.
5. Set up tipping. Tip prompts, tip adjust on pre-auths, tip pooling or tip-out rules if you use them.
6. Define staff roles and permissions. Who can void, comp, discount, open the drawer without a sale — and an audit trail for all of it.
7. Run live tests. Real orders, real payments, real voids and comps, on a quiet afternoon — with the kitchen actually cooking.
8. Drill the offline scenario. Unplug the internet: what still works? Offline card behaviour varies by platform — confirm yours in writing with the vendor rather than assuming.
9. Set the printer fallback. If a KDS screen dies, tickets must route to the kitchen printer automatically or with one tap — decided and tested before opening.
10. Run a mock service. Staff as guests, full menu, real timing. You'll find three routing mistakes and one menu error — that's the point.
11. Cut over on a slow day. Tuesday lunch, not Friday dinner. Keep the old system accessible for one week as a reference, not a crutch.

5. Install guides

6. Troubleshooting

7. Hardware options reference

Real models only. Links were verified to resolve on 2026-09-28; confirm stock, pricing, and compatibility with your POS platform before buying.

8. PCI scope callout

9. Integrations you may also need

Each integration is separated on purpose: what it is, why you'd want it, what it costs (real published pricing or "quote-based"), and an illustrative ROI sketch — a hypothetical example to show the math, not a promise. Prices below were observed in research during September 2026 and change often — verify current pricing before buying.

10. Everything integrated — full operational system map

11. What normal looks like

Honest benchmarks, honestly labeled. Restaurant IT-spend figures come from dated industry studies plus modern framing — treat the range below as a working range, not a quote: roughly 1–3% of revenue for independent full-service restaurants. Card processing research gives a broad 2.0–3.5% transaction-cost range, and a published 2025 figure put full-service margins around 2.8% — which is why the processing line deserves its own audit.

Labels: directional/working-range figures, not quotes. Processing fees sit on top of these numbers — see what yours actually are with the cost simulator.

12. How AI is changing this sector

Real products only. Costs are labeled by source type and date — verify current pricing before buying; none of these imply results.

• Slang.ai — AI phone answering for restaurants; around $399/location/month per third-party comparisons (observed September 2026).
• Maple — AI phone ordering; around $350/month monthly or $220/month annual per vendor pricing (observed September 2026).
• Loman — AI phone agent; $199/month per vendor blog (observed September 2026).
• Bite Buddy — AI ordering assistant; $1.50/order per vendor pricing (observed September 2026).
• MarginEdge — invoice and food-cost automation; around $350/month/location per industry reporting (observed September 2026).
• Restaurant365 — accounting and operations platform with automation features; quote-based pricing.
• 7shifts — scheduling with demand forecasting; around $45/month/location per AI-sector research (observed September 2026).
• Winnow — AI food-waste tracking; quote-based, usually enterprise or institutional scale.