WhatsApp for restaurants: reservations, delivery and digital menu with AI

A restaurant with two locations receives 80-150 WhatsApp messages per day on average: orders, reservations, menu questions, complaints, modifications. Handling that manually means either dedicating a person (who spends 60% of the time on repetitive questions) or ignoring messages and losing sales.

A well-implemented AI chatbot on WhatsApp solves the four critical flows of a restaurant business — reservations, delivery, menu/promotions and loyalty — without replacing the human team for cases that genuinely require it.

The typical restaurant stack on WhatsApp

Before talking about AI, it helps to map the tools already in play:

• Reservation systems: TheFork, OpenTable, Yumpingo, Google Reserve, manual Google Calendar.
• Direct delivery: WhatsApp orders with bank transfer, MercadoPago link, Stripe link.
• Third-party delivery: Rappi, PedidosYa, Uber Eats — where the brand pays 25-35% commission.
• POS: Fudo, Bistrosoft, Simphony, Maxirest.

AI on WhatsApp does not replace this stack — it connects to it. The real question: who answers the messages and how much does it cost?

Four flows to automate

1. Reservations with auto-confirmation

The customer writes "I want a table for 4 on Saturday at 9 PM". The AI:

1. Asks which location (if there are several).
2. Verifies availability in real time against the reservation system.
3. Confirms the booking, sends the address with Google Maps link, saves the customer in CRM.
4. 24h and 2h before, sends reminders with Confirm or Cancel buttons.

Typical result: going from 30-40% no-show reservations to under 10%.

2. Delivery orders with integrated checkout

End-to-end flow:

1. Customer opens a conversation with the brand.
2. AI shows the menu (rich text or WhatsApp Business catalog).
3. Customer builds the order in natural language — "one pepperoni pizza and two cokes".
4. AI confirms the order, calculates total with delivery fee, sends MercadoPago/Stripe payment link.
5. Once paid, the order enters the POS automatically.
6. Status update when the order leaves the kitchen with estimated arrival time.

Compared to Rappi/PedidosYa, this flow eliminates 25-35% commission per sale.

3. Menu and promotions

The most repeated questions are always the same:

• "Do you have gluten-free options?"
• "Until what time do you deliver?"
• "How much is the Wednesday deal?"
• "Do you take reservations for large groups?"

A well-trained AI answers them in seconds, 24/7, with up-to-date prices and hours from a single source (Google Sheet, custom CMS, or POS database).

4. Loyalty and campaigns

Once the customer is in CRM, you can trigger:

• Segmented campaigns by past consumption ("It's been 30 days since your last visit, here's 15% off").
• Birthday reminders with a perk.
• New menu launches only to customers who already ordered similar categories.

WhatsApp has a 95%+ open rate vs 18-22% for email — the ROI difference is huge once the database has consented to receive messages.

Multi-location without losing coherence

The most common mistake at multi-location restaurants: each location runs its own WhatsApp number and gives different answers. A centralized AI with location detection by geolocation or initial question solves this:

• A single brand-level number.
• AI detects or asks the location before any operational action.
• Each location only sees messages routed to it.
• Comparable metrics across locations.

Critical integrations

System	What for
TheFork / OpenTable	Reservations with real availability
Fudo / Bistrosoft	Push orders into the POS
MercadoPago / Stripe	Dynamic payment links
Google Sheets / Notion	Menu catalog the team can edit
Meta Lead Ads	Campaigns: delivery → conversation

Metrics management cares about

A restaurant manager with WhatsApp automation watches weekly:

• Conversion rate from message to booking or order: 35-55% is a good range.
• Average ticket: WhatsApp vs third-party delivery: typically 8-15% higher on WhatsApp due to guided upsell.
• No-shows on reservations: < 10% with automated reminders.
• Complaints resolved without human escalation: 60-75% if AI is well-trained on change/refund policies.

How to start without breaking anything

Suggested 10-day rollout:

• Days 1-2: register WhatsApp Business API and verify the number with Meta.
• Days 3-4: load menu, hours, locations and delivery policies into a single editable source.
• Days 5-6: integrate reservation system and POS.
• Days 7-8: configure confirmation, reminder and promo templates.
• Day 9: pilot with one shift or one location.
• Day 10: full rollout + weekly dashboard.

Next steps

If your restaurant is losing reservations and sales because no one answers WhatsApp in real time, see how Keebai handles reservations, delivery and menu with AI or book a demo and we'll set up a pilot with your real menu in under two weeks.