FOR RESTAURANTS

Train your restaurant into AI.

Every party-size, allergen, and dietary ask handled before the host can wipe their hands. A specialized AI trained on your menu, your reservation rules, and your house tone — answering on the channels your guests already use, not a third-party widget.

Start a trial →See pricing

What it knows.

– your menu — dishes, prices, sourcing notes, what is in the weeds tonight
– your allergens — every preparation, every shared surface, the cross-contamination story by dish
– your hours, reservation rules, and party policies — minimum spend, deposit, prix fixe windows, dog-friendly patio
– your dietary swap and chef-modification policy — what the kitchen will do, what they will not
– your house tone — how you greet a regular vs. a first-timer, what you say when a table is running late
– your special-event lead-in — anniversaries, birthdays, the table by the window, the cake-cutting policy

Where it shows up.

– SMS reservations — the Friday-night text asking if you have a 4-top in the next thirty
– website chat — the allergen question on the menu page
– voice for after-hours missed calls — the guest who tried at 10pm and got the answering service
– internal POS-side question handling — "does the salmon have soy?" from a server's tablet
– takeout-confirmation drafts — the "your order is ready in twelve minutes" SMS in your voice
– special-event intake — anniversary, birthday, private buyouts logged for the GM by morning

A REAL FRIDAY

What a trained restaurant AI actually says.

It is 8:04pm on a Friday. The host is in the weeds. A guest texts the main line. The AI checks the floor, offers the next slot, books on confirmation.

8:04pm Fri · Guest

any 4-tops in the next 30? my sister just landed early

8:04pm · Trained AI

Tight tonight — our next 4-top is 8:45pm, ninety-minute table. I can also seat you at the bar in fifteen if you are OK with high-tops; the full menu is available there. Which works?

8:05pm · Guest

8:45 works. one of us is GF and one nut allergy fyi

8:06pm · Trained AI

Booked — 4-top at 8:45 under your number. I am flagging both allergies for the kitchen now. About half the menu is naturally gluten-free; the tree-nut allergy needs the pesto and the romesco swapped, and Chef will confirm the pasta water station with your server when you sit. Anything I should pass to the kitchen tonight — birthday, anniversary, anything to make a fuss over?

8:06pm · Guest

nope just hungry, see you at 8:45

Dialogue is illustrative. Table availability, allergen tags, and the kitchen flag are wired to your reservation system once trained.

Before and after.

Before Gpodz

– the host bounces between the floor, the phone, and the chat widget; one of the three drops
– allergen questions on the website route to a generic answer the chef did not write
– after-hours missed calls go to voicemail and convert at half the rate
– special-event intake lives in three different inboxes; some never get a callback

After Gpodz

– SMS, chat, voice, takeout, and special-event intake all answered in your house tone, around the clock
– allergen answers come from your chef's playbook, not a default disclaimer
– after-hours missed calls get a same-minute reply with the next available slot
– events logged in one place, qualified for the GM's morning pass

EARLY OUTCOMES

What we expect to measure.

Real numbers, once design-partner runs land. Placeholders below are explicit; we publish measured values, not projections.

– missed-call rebooking conversion: TODO % — measured against the same restaurant's prior 30-day baseline
– host hours recovered: TODO hours/week — replacing the SMS + chat + voicemail rota during service
– allergen-question first-touch time: TODO (target: under a minute) — Sprint 005+ design-partner data
– special-event lead-to-booking rate: TODO % — paired pre/post baselines per location

Numbers will be replaced with measured values from design-partner runs once agreements sign and the training jobs complete.

An AI that has read your menu beats a generic one every time.

A generic AI does not know your menu, your chef's line on substitutions, or the way you handle a missed reservation from a regular. It writes a generic allergen disclaimer that is no use to anyone. A trained AI quotes the dish your kitchen is plating tonight, names your cross-contamination story for the tree-nut question, and texts back in your house tone — warm to regulars, brisk on a Friday at eight. The voice is yours because the trained intelligence learned it from your transcripts, not a default. It runs on shared infrastructure for a small fraction of frontier rates at SMB volumes.

Trust, quietly.

Three guarantees underneath every reply. The mechanics live on the trust page.

– Eval-gated replies. If the AI is unsure about an allergen edge case or a same-day private buyout, it routes to your GM instead of inventing.
– Verifiable billing. Every charge is Ed25519-signed and verifiable offline. Designed to start at Ready.
– Your data stays yours. Guest reservations, allergen profiles, and your house playbook never train Gpodz platform models, never get pooled across tenants, never get sold downstream.

Read the trust page →

Same shape, next door.

Hospitality, recurring guests, voice-matters, missed-call rescue — the same pattern shows up in the businesses around yours.

– Hotels — front-desk SMS, late-night booking questions, group-inquiry intake in your property voice.
– Medical spas — booking SMS, intake screening, contraindication checks in the front-desk tone.
– Contractors — SMS estimates and missed-call rescue trained on your pricing brackets.

Back to verticals →

Train the AI that knows your menu.

Start with one location. Ship it across every channel. Iterate from real guest texts.

Start a trial →See pricing