The AI voice agent that takes reservations
while service is in full swing.
Phone rings during the dinner rush. The host can't take it — they're seating a four-top. The reservation goes to voicemail. The customer books elsewhere. We install the AI that takes every reservation 24/7, handles large-group requests, runs the waitlist, and confirms takeout — in your brand voice, integrated with OpenTable / Resy / Toast. Live in 5 days. From $600/mo.
What the open loop is costing you.
The conservative numbers below are pulled from luup's audit data and published industry research for restaurant. Every leaky channel here is a competitor's win.
- · OpenTable industry data 2024: 40% of restaurant calls go unanswered during service.
- · Toast industry report: average reservation lifetime value $148 (covers + drinks).
- · luup audit of 6 mid-market restaurants, Q1 2026.
Wired into the tools you already run.
OpenTable + Resy have direct API access for restaurant partners — bookings write straight to the floor plan. SevenRooms uses webhooks. For takeout (Toast / Square), the AI captures the order, sends to the kitchen printer via the POS API, and texts the customer pickup time.
Four playbooks shipping in week one.
Each playbook is shippable within live in 5 days of kickoff. Most restaurant clients start with the first two and add the rest in weeks 2–4.
Reservation booking 24/7
Caller wants a 7pm Saturday for 4 — AI checks the floor plan, offers nearest slots if 7pm is full, books, sends SMS confirmation with arrival instructions (parking, dress code, special menu nights).
+25-40% reservations / weekLarge group + private events
Group of 10+ triggers a different script: AI captures size, date window, dietary requirements, occasion, and either books directly into the private dining slots or routes to the events manager with a structured intake.
3× large-group inbound captureWaitlist management
When the restaurant is full, AI offers the waitlist with an honest estimated wait. Sends an SMS when the table is 10 minutes from ready. Caller can confirm or release the slot — releases auto-roll to the next person.
~30% more covers / busy nightTakeout order capture
Caller wants takeout — AI reads the menu by section, takes the order, captures dietary notes, calculates total, sends SMS payment link, fires the order to the kitchen printer. Pickup time set to current kitchen capacity.
Frees ~6 hrs/week of host timeReal numbers from a real operator.
An 80-seat restaurant we benchmarked recovered 60+ reservations / week
What it costs, without the dance.
Restaurant deployments run $600/month for reservations only, $1,000/month with waitlist + takeout, $1,400/month for restaurant groups (multi-location with shared script + per-location floor plans). Setup is $1,800-$3,500.
Five things restaurant operators ask first.
The questions every restaurant prospect asks on the first call. Answers in writing so you can decide before booking.
Can the AI handle our specific menu and allergens?
Yes. We load your full menu, allergen tags (gluten, nut, dairy, shellfish, vegan), and any seasonal specials during week-1 setup. The AI can answer ingredient questions, flag allergen-sensitive bookings to the kitchen, and refuse to book if the requested time conflicts with a private event.
Will it speak in our brand voice?
Yes. ElevenLabs voice cloning gives you a custom AI voice — we usually clone the owner or a senior staff member with their permission. Tone is tuned per-restaurant: casual neighbourhood spot vs. fine-dining vs. members-only club all sound different.
How does it handle the waitlist on a busy Friday?
Same way a senior host would: gives an honest wait estimate based on current turn time, captures the party's name + phone, sends SMS when the table is 10 minutes out. Caller can confirm or release — releases auto-roll to the next person, no human intervention.
What about complaints or VIPs?
Two hard handoff triggers: complaint sentiment ("the food was terrible last time") and VIP recognition (configurable list of phone numbers that always go to the GM). Handoff routes to the on-shift manager's mobile via Twilio in under 10 seconds.
What's the per-call cost during a busy service?
Vapi + ElevenLabs + Twilio variable cost runs ~$0.18-$0.25 per call (3-min average). For a busy 200-call night that's ~$45 in variable cost on top of the monthly fee. Pass-through at vendor rate, no markup. We cap usage with budget alerts so it can't blow past your forecast.
What we’re writing about restaurant.
Latest from the luup blog, tagged for restaurant operators. Numbers, scripts, real benchmarks — everything we’d send a prospect on the first call.
Other verticals we ship voice agent for.
Same Vapi + ElevenLabs + Twilio voice agents, tuned to industry-specific needs. Same deploy cadence. Same guarantees.
Ready for restaurants
voice agent?
Book a 30-min strategy call. We’ll map your current setup, show you the system in action, and quote the deploy. If it’s not a fit, we’ll tell you on the call.