Verneek Launches Generative AI Platform to Help Food Consumers
Today Verneek, a New York-based AI startup, debuted its first product, Quin Shopping AI. The product is the first to use the company’s proprietary AI platform called One Quin.
The company, founded by the husband-and-wife team of Omid Bakhshandeh and Nasrin Mostafazadeh, spent the last two years developing the One Quin AI engine, which Mostafazadeh describes as an ‘AI platform for consumer experience.’
“What we have done is that we have created a system with many integrated modules for different transformer technologies or non-transformer technologies that are trained to answer incoming questions,” said Mostafazadeh.
According to Mostafazadeh, Shopping AI was trained on anonymously aggregated consumer query data collected by the company’s first partners (which they say they cannot disclose at this time) and automatically generated data sets based on these consumer queries.
Mostafazadeh said One Quin Shopping AI is different from other artificial intelligence systems, such as ChatGPT, because it is specifically targeted at a specific application area of the consumer shopping experience.
“Quin one is AI and selected information in the box, while the likes of ChatGPT is general AI where information is not limited.”
Another benefit of this specific focus is that, according to Mostafazadeh, their product will not suffer from the optical illusions that plague manufacturing AI systems. General-purpose generator AIs like ChatGPT will sometimes produce answers that, while seemingly believable, may be factually incorrect or nonsensical. In contrast, the One Quin is fixed with certain parameters within a closed subject set and designed in such a way that it produces reliable answers.
“We spent the last two years reducing that (visualization),” Mostafazadeh said. “What’s very different from what we’ve created is that One Quin sits on top of the data. So it doesn’t produce it externally. Instead, with very complex internal mechanisms, it points to the data that sits on top of it.”
Mostafazadeh said that because the One Quin engine points to specific data, it can answer specific questions that match the parameters that consumers use when searching for a product. For example, let’s say a customer has a question about a food or nutritional product that fits a certain price range. In that case, One Quin can access this data and generate a corresponding response specific to the merchant’s product range.
“What Quin can do, for example, is answer a question like ‘what healthy meal can I buy for my kids that costs less than $5?'” Mostafazadeh said.
I asked Mostafazadeh how his AI can determine whether a product meets criteria such as health, which can sometimes be arbitrary. He told me they did something like a “health score” based on nutrition research. In some vague terms, he told me that the system is designed to support responses with data points that they believe serve as a good proxy.
“To get the taste, Quin is based on the ratio that things have,” Mostafazadeh said.
In time, however, Mostafazadeh says they could develop a way to get the product to taste more accurately. However, one challenge in that, for now at least, is that the system is currently designed to answer questions without information about the consumer.
“Right now, we’ve decided to make the barrier to entry zero. We don’t even ask buyers to come in. We don’t track them, so it’s a blank slate.”
That could change, says Mostafazadeh, who admits that adding a personal shopper mode can be very powerful.
“We’d like to know if you’re vegan unless you tell me you don’t eat meat in your question. I’d like to know if you hate cilantro because it tastes like soap, and by default, I’ll show you all the recipes that don’t have cilantro in it.”
Mostafazadeh said another advantage of Open Quin is that it can sit on top of any computing engine, whether it’s Microsoft Azure, AWS, Google Cloud, or in-store edge computing architecture. He said this makes it more affordable than other artificial intelligence systems and gives vendors — who can be very specific about which cloud or computing infrastructure they tie to — more flexibility.
“You probably know that retailers don’t like AWS (Amazon’s cloud). They don’t want anything in their country that touches anything Amazon.”
Mostafazadeh said Quin Shopping AI can be deployed using unique user interfaces. For example, he said retailers can use it in an app, on a website, through a chatbot, or at a consumer kiosk.
The company raised $4.2 million in pre-seed funding, and its website went live today.
