Certus AI Wants to Answer Every Restaurant Phone Call That Nobody Picks Up

The Macro: The Phone Is Ringing and Nobody Is Picking Up

Here is the dumbest revenue leak in the restaurant industry: missed phone calls. A single-location restaurant gets somewhere between 50 and 200 calls a day depending on the concept and market. During the lunch rush, nobody is answering. During the dinner rush, the host is juggling walk-ins and the phone rings six times and goes to voicemail. After 9 PM, the phone is off.

Every one of those missed calls is a potential order, reservation, catering inquiry, or event booking. The industry estimate is that a typical restaurant loses $100,000 per year from calls that do not get answered. I have talked to operators who think that number is conservative.

This is not a new observation. The restaurant tech space has been trying to solve it for years. Popmenu does AI phone answering. Slang.ai handles voice automation for restaurants. Maitre-D AI and Tock have reservation-focused voice products. The phone-to-order pipeline is a known problem with a growing number of competitors.

But the solutions that exist tend to fall into two camps. The first camp is basic IVR trees that sound like robots and frustrate callers into hanging up. The second camp is more sophisticated AI voice products that work well for reservations but cannot actually take a food order, process modifications, apply coupons, or push the ticket into the POS. They answer the phone. They do not close the sale.

The restaurant POS ecosystem is also fragmented in a way that makes integration genuinely difficult. Toast, Square, Clover, Skytab, SpotOn, Aloha, MICROS. Each one has different APIs, different menu structures, different modifier logic. A voice agent that works with Toast but not Clover is only useful to a fraction of the market.

The Micro: Three Founders, Multilingual Agents, and POS Integrations That Actually Work

Certus AI is building a voice agent that replaces the restaurant phone line entirely. The product answers every call 24/7, takes food orders with full modifier support, handles reservations and event inquiries, suggests upsells, and pushes completed orders directly into the POS.

Gurveer Singh is the co-founder and CEO. Adam Gamieldien is the co-founder and CRO. Isaac Nichols is the co-founder and CTO. They are a three-person team out of San Francisco, part of Y Combinator’s Summer 2025 batch.

The POS integration story is the part that matters most. Certus works with Toast, Square, Clover, and Skytab, with more integrations in the pipeline. When the AI takes an order, it does not create a note for a human to process later. It builds the actual ticket in the POS, with modifiers, special instructions, and payment. The kitchen sees it like any other order.

The multilingual capability is a real differentiator. Certus handles calls in English and Spanish, and the team claims it manages a wide range of accents well. For restaurants in diverse markets, this is not a nice-to-have. It is table stakes.

Setup takes 48 hours and requires about 45 minutes of the restaurant’s time. That is fast. Most restaurant tech products require weeks of onboarding and training. The low friction matters because restaurant operators are busy people who have been burned by software that promised easy setup and delivered a three-month implementation project.

The customer testimonials tell a consistent story. Art Ingemi praised the response speed and problem resolution. Charandeep Modi said the AI handles calls better than staff, which is a bold statement from someone who was initially skeptical. James Young called it “brilliant” for ensuring no calls are missed.

The smart upselling feature is interesting. The AI does not just take the order as stated. It suggests add-ons based on what the customer is ordering, the same way a trained counter person would ask if you want fries with that. The revenue impact from systematic upselling across every call could be significant.

The Verdict

I think Certus is well-positioned in a crowded space because they are solving the full problem, not half of it. Taking a reservation is easy. Taking a food order with double pepperoni, no mushrooms, extra sauce on the side, and a coupon code, then pushing that into Toast as a complete ticket, is hard. That is where Certus is focused.

The risk is competition. Popmenu has scale and a broader restaurant marketing platform. Slang.ai has funding and brand recognition. If any of the POS companies decide to build native voice ordering, that changes the landscape overnight. Toast in particular has the technical capability and the distribution to make a voice product a default feature rather than a third-party add-on.

The other risk is call quality edge cases. Voice AI sounds great in demos. It sounds less great when a caller has a thick accent, bad cell reception, background noise from a construction site, and wants to order for twelve people with three separate payment methods. The gap between demo and production in voice AI is real, and it only takes a few bad experiences to lose a restaurant’s trust.

In 30 days, I want to see average order value comparisons between AI-handled calls and human-handled calls. If the upselling works, that number should be meaningfully higher. In 60 days, I want to know the call completion rate. How many callers who start an order with the AI actually finish it without asking for a human? In 90 days, the question is multi-location expansion. Can Certus onboard a 20-location franchise in a week? That is where the revenue scales and where the operational complexity gets real.