What is artificial intelligence?
Many myths have been created around AI. The most popular of these is confusing artificial intelligence with artificial consciousness. What is referred to as AI has no feelings, no sense of existence and, above all, lacks consciousness. Artificial intelligence is much more ordinary. AI is a mixture of advanced mathematics and computer science, which allows solving complex problems requiring non-obvious solutions and adjustments to new situations.
So what makes it different from the electronics you use every day? A computer algorithm needs to be written by a human – the author must first figure out how to solve the problem and then develop the appropriate program for the computer. In machine learning, the situation is different. A human writes an algorithm that determines how the robot should learn – showing it the data and the goal. Based on these components, the computer is able to independently find relationships in the data to achieve the objectives, for example, of allocating orders to drivers so that the orders are delivered as quickly as possible.
Can artificial intelligence predict food delivery times?
As shown by Papu.io’s report on customer satisfaction with food delivery service, the most important factor indicated by 40% of respondents is the conformity of the driver’s arrival time with the so-called promised time. This task is particularly challenging for a human. It requires analyzing and coordinating the work of the kitchen and drivers. Determining delivery times without accurate calculations is as difficult as coming up with a formula for calculating delivery times and entering it into a computer.
Deciding on a meal delivery time is much easier with the assistance of artificial intelligence. Why is that? The Papu.io system has access to detailed information on several million completed orders. Taking into consideration more than 20 different characteristics describing the kitchen facilities, the drivers and the town, the Papu.io engineers built and taught artificial neural networks how to predict the exact delivery time of an order. In machine learning, such a task – i.e. predicting an exact value, for example, 21.37 minutes – is called regression.
How does artificial intelligence distribute tasks in the kitchen?
One problem that comes up time and again in the kitchen is the selection of orders that need to be completed first. If a dish is made too early, it will have time to cool down before the driver arrives to collect it. However, if there is a delay, the driver waiting for the dish will be generating losses. Here again, as in every non-obvious and difficult case, AI comes to the rescue. Making use of Papu KDS app, it is possible to build a smart system for organizing the work of the kitchen staff and drivers. Its responsibility is to inform the employees which orders must be started first, which orders are to be collected and when it is possible to take a break.
The Papu.io team has engaged exceptional resources in developing this solution. The system takes advantage of the so-called reinforcement learning technologies which are used when driving autonomous cars, winning against chess champions and beating human players in computer games. From the point of view of artificial intelligence, it plays a game with the aim of delivering orders to customers as quickly as possible.
Can a restaurant enjoy better ratings thanks to artificial intelligence?
People say: “if you want to learn what competition is, open a restaurant”. The choice of restaurants is enormous, especially in big cities visited by thousands of new people every single day. This is where the so-called social proof, i.e. customer reviews and ratings, is particularly crucial. Social proof corresponds to restaurant ratings on Google, Pyszne.pl or other portals for ordering food which encourage new customers to place an order. Papu.io’s customer satisfaction module was developed to monitor customer satisfaction with their orders and share positive feedback with other consumers. Unfortunately, restaurants very seldomly have thousands of positive online reviews, so they cannot afford to have even several negative comments posted about them. At present, the Papu.io team is working on something very special. If we make a mess of the delivery service, the customer will likely share this fact online. But what if you knew in advance which order would be a dud, allowing you to prepare yourself to appease your customer’s wrath? Based on the data collected from the customer satisfaction module, Papu.io is working on a solution that will use artificial intelligence to predict the so-called “problematic orders” as early as at the order preparation stage. Informed of potential problems in advance by means of an automatic notification, the driver will be able to add, for example, a complimentary drink to the order in question.
No computer will ever replace a helpful waiter, a talented pizzaiolo or the restaurateur’s pride in building a close-knit team. However, as the above examples have shown, waiters, chefs and drivers continuously face serious challenges in their daily work. Fortunately, low-cost and effective AI solutions are on hand to help us overcome all these issues. The kitchen staff will no longer be required to focus on complex calculations, the customers will receive their hot food on time, and the restaurateurs will be able to stop worrying about what is happening inside their business premises.