The Phone Is Ringing. Nobody Is Picking Up.

The phone rings during dinner rush. Nobody picks up. The caller hangs up and orders from a delivery app instead. Most restaurants treat this as an unfortunate moment — the equivalent of a walk-in who looked at the menu and left. It is not. It is worse.

What a missed call actually costs

A walk-in who leaves has at least seen your dining room. They have a sense of what the experience is. They may come back. A caller who gets voicemail is simply gone — and that dinner went somewhere else.

The math is direct. A family of four who wants to order dinner for pickup represents roughly $40–80 in revenue, depending on the restaurant. If your phone line is busy or unanswered for four hours a night and that happens twice, you have lost the equivalent of a full table. Every night.

For a restaurant running 60 seats with an average ticket of $55, three missed calls during a dinner service is $165–330 in that one service period — before the cost of the staff time spent answering calls that did connect.

The monthly leakage — three hypothetical tiers

The cost of missed calls compounds quietly because it never appears in a report. It is not tracked in any POS dashboard, not captured in any delivery app metric. It is simply the revenue that did not arrive because the line was busy or the voicemail was never checked.

At 10 missed calls per week at an average ticket of $50: $500/week in lost orders, $2,000/month, $24,000/year — before accounting for the callers who would have become regulars.

At 20 missed calls per week at an average ticket of $60: $1,200/week in lost orders, $4,800/month, $57,600/year.

At 40 missed calls per week at an average ticket of $70: $2,800/week in lost orders, $11,200/month, $134,400/year.

These tiers are illustrations, not measured data. Missed calls never appear in any metric — until the owner notices the delivery app orders are higher than expected and the walk-in volume is lower than it should be.

The multilingual gap — where most restaurants lose calls they do not track

For a significant portion of North American restaurants, the missed-call problem is also a language problem. At a Chinese restaurant where the next caller might be speaking Mandarin, Cantonese, or English — sometimes switching between all three inside one conversation — the person who answers the phone must be fluent in all three.

King Do Seafood in Surrey, BC answers calls in two languages. Lao Cai on Marine Drive answers in English and 中文. Yuan's Garden in Richmond takes calls in Mandarin, Cantonese, and English. In each case, the language mix is not an edge case — it is the baseline. And the people who could answer the phone in all three languages are often the people who are also cooking.

The missed call at a multilingual restaurant is not just a call that was not answered — it is a call that required a specific bilingual capacity to answer, which no one on shift may have had at that moment.

The commission trap — and how it compounds the problem

Many restaurants responded to the missed-call problem by directing customers to third-party delivery apps. It works — but the cost compounds. Third-party platforms charge 20–30% commission per order. For a $60 ticket, that is $12–18 per order going to the platform, not the restaurant.

The alternative — paying someone to answer the phone — has its own costs. During a dinner rush, a dedicated phone person is the most expensive person in the building per minute, and they are doing a job a machine can handle. Add the staff time spent taking calls that could have been orders, and the real cost of manual phone handling is higher than it appears.

Third-party apps also introduce a second cost that is invisible until it is too late: they own the customer relationship. Every order taken through the app belongs to the platform. The restaurant has no direct access to the caller's phone number, email, or order history. A diner who discovers the restaurant through a delivery app and orders through it will continue to reorder through that app unless something actively changes that pattern. The phone channel — the restaurant's own direct line — is the channel that breaks that cycle.

What the phone channel captures that delivery apps never do

A delivery app knows what someone ordered. It does not know who called to ask whether the restaurant had parking, who asked about the private room for a birthday party, or who hung up because the wait was too long. The phone captures intent that never became a transaction — and that intent is information.

The caller who asked about a large party and hung up without ordering is a missed catering lead. The regular who called to check hours and then drove in is a visit that an app could not generate. The first-time caller who got voicemail and ordered from a delivery app instead is a customer the restaurant will probably never recover. The phone channel, when it is answered, builds a customer profile the restaurant owns — not the platform.

What AI phone answering changes

An AI phone host answers on the first ring, every time, 24 hours a day. It speaks the languages the restaurant's callers use — English, Cantonese, and Mandarin — detected within one or two words and switched mid-call if the caller switches. Response latency is under 800ms, so the call feels natural.

At King Do Seafood, the AI has taken over 100 phone orders since deployment. Their owner Darrow put it this way: "When we get busy, we don't have to worry about missing calls from our customers any more." The order was captured, confirmed on the call, and sent to the kitchen before the caller hung up.

The AI captures the full order on the call — items, modifiers, pickup time — and sends it to the kitchen board. It captures the caller's phone number for the customer profile. It recognizes returning callers and greets them by name. Every call that comes in is a data point, not a missed opportunity.

The pricing comparison at three volumes

The founding base fee is CAD $199 per month per location for the first 100 partner restaurants, locked while continuously subscribed; the standard base fee is CAD $399. Starter is CAD $99/month/location with 200 AI call minutes, then CAD $0.35/min. Standard, including its founding offer, includes 1,000 AI call minutes per location per month, then CAD $0.20/min. Pro is CAD $599/month/location with 3,000 minutes, then CAD $0.15/min. Online and QR orders use no call minutes. Taxes extra. Zero order commission and no setup fee. Includes an AI phone ordering agent, a restaurant website, online ordering and reservations, QR ordering, and a unified management dashboard.

For an illustrative four-week month with 20 orders a week at CAD $55 each, order revenue is CAD $4,400. A 20% commission would be CAD $880. DineCall takes no commission; the founding bill is CAD $199 before tax if AI call time stays within 1,000 minutes, with CAD $0.20 per additional minute.

At 40 orders a week at CAD $70 each over four weeks, revenue is CAD $11,200. A 20% commission would be CAD $2,240. The DineCall founding base fee remains CAD $199 per location, but total fees depend on AI call minutes: usage beyond 1,000 minutes costs CAD $0.20 per minute, before tax.

Every plan also includes a restaurant website — designed, built, and hosted. Most restaurants pay separately for that. Third-party delivery apps charge for a marketplace listing. The website DineCall includes belongs to the restaurant.

What does not change

The phone number stays the same. The POS stays the same. The kitchen workflow stays the same. DineCall answers the line you already publish. There is no new hardware and no new app for staff to learn.

Calls that need a person — private rooms, large banquets, catering — go directly to the team. The AI handles the straightforward orders and bookings end-to-end, so the team is interrupted only when it matters.

The question is not whether the phone is worth answering

Every restaurant knows their phone is worth answering. The question is whether the current solution — staff time, missed calls, third-party commissions, or voicemail — is the right answer. The math usually says it is not.