3
9 Comments

I'm 16 and I just launched a voice AI that answers restaurant phone calls

I’m 16, based in France, and I’ve spent the last few months building Linéa AI — a voice AI that answers restaurant phone calls 24/7, takes orders in natural language, and sends them straight to the kitchen printer.

The idea came from noticing how many independent restaurants (pizzerias, kebab shops, fast food places) lose orders every day simply because they can’t answer the phone during rush hour. Existing solutions were either generic voice assistants that don’t understand food orders, or too expensive for small restaurants.

I’m currently onboarding my first restaurants in the Paris area. Happy to answer any questions about the tech stack, going solo at 16, or anything else — would love feedback from this community.

posted toAvatar for product Linéa AI
Linéa AI
  1. 2

    What stands out to me is less the tech and more the sales motion — convincing an independent restaurant owner to trust an AI with their phone line is a hard sell even without the "I'm 16" factor. How are you actually getting the first restaurants to say yes? Cold walk-ins, personal network, free trial period? That part seems just as hard to crack as the voice tech itself.

    1. 1

      Honestly, right now it’s almost entirely cold walk-ins — I go into pizzerias/kebab shops in person (I’m based in France) and offer to demo the assistant live, right there, calling their own number with their actual menu, rather than asking permission first and pitching abstractly. That live demo does most of the convincing.

      I’ve also started sourcing leads from BODACC — a French public government register of business filings, including recent restaurant ownership changes. New owners tend to be more open to changing how things run than someone who’s had the same setup for 10 years. I cross-reference that with public professional directories (like PagesBlanches) to find the actual manager’s name and listed business contact before reaching out, rather than just calling the generic restaurant line.

      No free trial yet, that’s honestly a gap I know I have compared to some competitors. My most advanced deal right now is still just a scheduled call, no signed client yet, so I don’t want to oversell how solved this part is. Honestly, the sales/outreach side has turned out to be at least as hard as building the voice tech itself — maybe harder, since the tech I can just keep improving alone, while getting a stranger to trust you with their phone line requires them to say yes.

      1. 2

        Love the BODACC angle — sourcing leads from ownership changes instead of cold-calling blind is smart. The live-demo-on-their-own-menu trick is gold, might steal that idea for Ehugapy. Good luck landing that first signed client!

        1. 1

          Thanks, and good luck with Ehugapy too! Feel free to steal the live-demo trick, it’s honestly the single thing that’s worked best for me so far.

          1. 2

            Thanks! I’ve just sent it over.

            Looking forward to hearing your thoughts whenever you have a chance.

  2. 2

    The strongest part is the wedge: missed phone orders during rush hour are a direct, measurable revenue problem for restaurants. The key question now is whether Linéa can handle the messy edge cases of real food orders reliably enough to earn a restaurant's trust.

    1. 1

      That’s exactly the right question, and I don’t have a fully proven answer yet since I haven’t signed a paying client. What I can say: the assistant is configured with the restaurant’s actual menu and real options (sizes, sauces, supplements) rather than a generic template.

      Before ever testing it live with a restaurant owner, I stress-tested it myself with a bunch of fictional menus, throwing tons of edge cases and messy orders at it deliberately. It did make mistakes early on — I went through and fixed each one as I found it. Since then, testing it regularly, it hasn’t made an error in a while.

      I also test it live with the owner during the demo so they can throw real orders at it themselves before deciding anything. If it hits something outside what’s configured, it transfers to a human instead of guessing. But the honest answer is: reliability at scale, across many different restaurants and real customers, is the thing I still need to prove and not claim.

      1. 2

        That makes sense. I’d be interested in continuing the conversation a bit more—what’s the best email to reach you at?

        1. 1

          contact.lineaai@gmail.com — happy to keep chatting here too if that’s easier for you.