About US$25bn of food goes to waste every year, including an estimated 4–10% of the food purchased by restaurants thrown away before it even gets cooked. Barcelona-based SaaS provider Proppos FastPay is tackling the problems that lead to such food wastage in restaurants, self-service cafes and office canteens, caused by operational inefficiency during rush hours when customer turnout spikes.
Powered by computer vision technology that recognizes food, Proppos' self-checkout machine improves operational efficiency and customer experience in restaurants. Restaurants that integrate Proppos' self-checkout solution into their enterprise resource planning (ERP) can speed up their payment queues, especially during rush hours, while reducing manpower cost since the process is fully-automated.
“Our technology is capable of recognizing food in less than one second and, depending on the type of payment the customer chooses, the self-checkout time is kept to between 5 and 10 seconds,” Proppos CEO and co-founder Nil Salomó Bellavista told CompassList at the 2019 IOT Solutions World Congress in Barcelona.
Further, the startup's AI-based software generates predictive consumption models using the data collected by the machines in order to reduce food wastage. Studying the issue of food wastage, Proppos's founders realized that customers tend to be unaware of their nutritional requirements and often over-order the amount of food they need. Restaurants, on the other hand, are unable to predict customer demand for the various dishes on their menus and thus cannot optimize their purchasing of ingredients accordingly.
Applying AI to the big data collected by Proppos' software can generate accurate predictions based on consumption models to optimize the use of resources, resulting in dramatic improvements at the operational level. Moreover, estimations can be enriched and refined over time as more data and insights are collected. Studies show that there is an economic incentive for minimizing food waste since it results in a 2–6% cost reduction.
Validation and scalability
Proppos started testing its product at the beginning of 2019 via a pilot with Nostrum, a food-takeout franchise with outlets across Spain. Pau Vaquero Pérez, a telecommunication engineer and long-time friend of Salomó's, accepted the latter's invitation to join Proppos as its CTO and co-founder. The two worked on MVP development, iterating and validating their technology via Nostrum's operations.
“My family owns Nostrum, and my brother and I manage one franchise. Having known this business for more than 20 years, I’m aware of the problem that this sector faces,” said Salomó, adding, "More than a pilot, Nostrum was a test and therefore we didn’t monetize through it.”
Between June and September this year, the startup started to implement its solution in three paid pilots with clients – whose names Salomó declined to reveal – located in the US, France and Spain. Proppos is currently looking to begin pilots in the Netherlands with global consulting, technology services and digital transformation company Capgemini and in Spain with a big football team that Salomó is not naming as yet.
He said, “We are happy with the clients we are working with. Although we are currently offering solutions at a reduced price, we are learning a lot in the process. And if the pilots work as expected, we have signed commercial agreements with these clients to guarantee a continuity of our collaborations in the long term.”
The pilots will validate Proppos' business model and pricing, preparing it for its official commercial launch in 2020. The company's revenue streams include a monthly license fee of about €500 per self-checkout machine as well as a setup fee of between €3,000 to €5,000 that includes the machine, which is fully-manufactured in Barcelona by a local hardware manufacturer and costs approximately €1,500 each.
“We are not trying to earn any margins from this [selling the machine],” said Salomó. He explained that “in the beginning we wanted to sell only our SaaS, but clients did not understand very well how the technology worked without seeing a machine functioning. That’s why we decided to sell the whole solution, including the hardware.”
During the setup phase, Proppos makes sure the self-checkout machine is capable of recognizing each food element on the plate and the dishes on the restaurants’ menus. To do so, it gets a huge amount of data from hundreds of photos of each food ingredientand uses a data augmentation process to complete the data set. According to Salomó, this process takes only a few hours, depending on the clients' requirements and menu options. He noted that since scalability is key for Proppos, future machine setups will be handed over to dedicated distributors and “they will be able to setup the machine autonomously, improving the scalability of our model dramatically”.
Reality check
The company is also looking to harness the huge amount of data that can be potentially collected by the self-checkout machines to provide nutritional insights and suggestions that will educate customers while giving restaurants consumption predictions so they can optimize their operational costs.
“We realized that our clients don't have clear information about their products which leads to bad operational results, like food wastage; they also have no idea about the quantity of food that has to be pre-prepared as they have no estimations of how much will be consumed,” said Salomó.
Proppos currently counts on a team of five and a high-powered group of advisors, who hold executive positions in Nostrum and supermarket chain Lidl and are serial entrepreneurs of technology-based companies. The startup recently received a €125,000 capital injection from its founders, 3Fs and local angel investors. It aims to close a second round of investments of between €500,000 and €1m by the end of this year, with the funding channeled towards reinforcing the team and improving its technology.
One of the biggest challenges that Proppos' founders have faced so far is with the food-recognition technology. Salomó recalled, “We had a pilot where the menu had spaghetti alla carbonara. One day, the chef changed the dish to spaghetti a la bolognesa and our technology didn’t recognize it because they are obviously two different things”.
The huge variety of food available and the fact that restaurants need the flexibility to change dishes on the menu proved to be a big challenge. Proppos' SaaS iteration now includes features that enable restaurants to make changes without interfering with the self-checkout machine's detection of every product sold to customers that is included in the system. Salomó concluded, “This is why clients' validation is essential to getting a new technology up and running at its full potential.”
For now, Proppos is focused on the Spanish market. Salomó said: “Although [FastPay is] applicable to retail, we will initially focus on restaurants, and self-service and corporate canteens … [in itself] already a big and complex industry.”
The company has begun developing a PROPPOS Free checkout, a Spanish version of AmazonGo for restaurants and retail stores, though. To finance this, the founders plan to tap Neotec Capital Riesgo, a Spanish-based fund launched by the European Investment Fund.
“Being the European AmazonGo would be amazing, there are many companies across the world like US, China, Israel introducing these type of solution in retails. But fir now, we want to focus on our segment to establish and grow our position in the market,” Salomó said.
