OrderAhead

Order Food from your CarPlay using A.I.

Comment

What inspired you

I worked for Apple and currently work for Chick-fil-A. (As a startup founder taking a day job) I ran into the problem of knowing there are only two apps to order food from CarPlay - Dominos Pizza and Chick-fil-A. I imagined "what if you could open an app in CarPlay and simply tell the app what to order. The idea is that a custom LLM using location data could query any database and provide ordering info. Additionally, we could could use AI to call local bodegas and restaurants without an app and ordering system and talk to a human and place an order in advance for our customer to pickup. We see a world where LLM's can allow for conversational ordering without the use of an app or interface - for a seamless user experience

What Order Ahead does

A four step process of voice ordering. STEP 1: Open app. On first screen place your order by speaking into app. The app will use an LLM to give you options based on your intent, location and its database. It will have full knowledge of mainstream restaurants menus and prices and will construct a total for you. STEP2: After you are satisfied press screen to got to next step for payments. Ideally since it is an apple based app - ApplePay is the first option of payment. After payment is selected, Our internal app will use n8n and the LLM to transmit the order for Automation and take the returned data and provide a summary of the order for the customer to walk into the store with. STEP 3: We add our own ordering number beginning with A.O. + plus a number plus a suffix. STEP4: Followup up and feedback. We query the user on pickup using geo-location and and acceptance prompts. We then bill the user upon successful pickup.

How we built it

I tried Bolt.new but could not get what I wanted. So I switched. I started with Manus for concept, Moved to Gemini for the PRD to prep the app idea then moved to Rork to build the final product in an attempt to submit a working app to TestFlight. I have now moved on to Cursor with Gemini CLI and am building payment API's and other rails in the hope of getting it to work. This is ongoing.

Challenges we ran into

  1. Protecting the IP related to the approach. We are entering an era of AI based apps where LLM's will help anyone make purchases. Our approach is first to market. It can function cross platform, being the first ubiquitous global voice ordering app. When paired with stable coin, pay by bank option it represents a fundamental change from Seamless or UberEats for seamless ordering.

    1. A Custom LLM - Perplexity is the best for natural language queries according to our concept - however it is expensive and subject to hallucinations. We envision a custom LLM built on top of Foursquare, Google Reviews and Yelp data to provide specific restaurant database and menu. This will allow for quicker response times.

*Apple Pay Integration and App Store Approval *- Our business model - A pay per use, voice ordering platform using a custom LLM will be a direct competitor to Apple's own Apple Intelligence for ordering. In testing - Apple Intelligence can already make suggestions and open and give directions to a restaurant. It does not allow you to place an order.

Live User Testing - Not possible until other hurdles are me. We are still very much at the idea stage.

Accomplishments that we're proud of

We were proud to simply have started. This calls attention to what we are trying to build. The next aim is to document on LinkedIn and build with support.

What we learned

OrderAhead seems simple, but it is incredibly hard and expensive. There are many roadblocks. To stay true to Apple's walled garden indeed provides a complex layer for the individual developer as with out additional supports getting the app to the App Store and into CarPlay seems impossible. We hold on to the belief that once we have an App into the App Store, then we can follow with one in the Google Play Store for Android Auto.

What's next for OrderAhead

We are at the very beginning. There is much to be done. Our immediate need is to incorporate and build out a small team. We also need to build our legal team and fintech lobbyists to get accepted in the App Store and Play Store. Google is about to provide voice enabled payments with Google Pay as per the government of India's requests. We anticipate a name change due to copyright issues with orderahead.org. Our road map includes building out modules with AI to order at restaurants without a digital presence. This involves using AI to call a bodega, query and place an order for pickup. We need to build or LLM, test it. Then we need to build language translation into our custom LLM (its in perplexity) and the logic models. We need to implement the AI customer service module using FIN ( http://fin.ai) and internal dashboard and reporting.

Shout out to the Tech Snack Community on Skool for your support. Explainer Video on idea with CarPlay demo. ![https://youtu.be/IzIRy1RrnOQ?feature=shared]

Built With

Share this project:

Updates