News

Meet TasteGPT, Tastewise’s AI chatbot helping companies predict the latest food trends

TasteGPT, a Generative AI chatbot, helps companies in the food and beverage industry anticipate future trends and use those insights for new product development.

In 2020 Ingredients Network caught up with Alon Chen , CEO and co-founder of Tastewise. At the time the company had just launched in the UK. Almost four years on, we followed up with Chen to understand how Tastewise, an AI-powered platform, has evolved its capabilities in predicting the way consumers order, cook, and eat.

Tastewise uses predictive analytics and forecasting to help companies anticipate future trends

Tastewise captures real-time data to provide insights into consumer behaviour. Using predictive analytics and forecasting, the platform helps companies anticipate future trends and make informed business decisions.

The forecasting and predicting trends process is twofold. The first step, having access to multiple data sources, is most important, Chen explains. “That allows you to understand the consumer. So we're using social data, traffic data to websites, and menu data. We are monitoring the menus of restaurants in every single restaurant in the countries we're present in, more than ten countries now. We are also using retail data.” The key here, according to Chen, when making predictions, is ensuring you’re looking at multiple data sources.

The second step is to take each of those data sources and ensure the company understands the trendlines. “What we are doing is basically breaking the trend into different segments. If we are looking at social data, is it happening across a nation, or is it a specific area? Is it across audiences? Or is it actually just one or two?”

Next, a regression is calculated to determine where the trend is present, for example, is it present in big chains only, or is it already independent in a specific region or zone? Chen says that knowing this information allows Tastewise to take very granular data and assess its accuracy in predicting trendlines.

“It is really, really important. One of the biggest value propositions or services we're giving to our customers is the ability to call ourselves predictive analytics. We're cutting-edge, but we're also explainable AI.”

Explainable AI, Chen says, is not about telling companies what the raw data shows about a future trend, it is also about building trust and helping them sell the idea internally, within their company.

“What we've discovered over the years that is as important as giving you a trend is helping you make decisions, move decisions, and get consensus in your respectful organisation, because this is very much needed.”

Quantitative data is one thing, but a company needs to be convinced about the data. To be convinced there needs to be a qualitative aspect to be able to explain what the data means to them, says Chen.

TasteGPT: “… even my mom can log on to Tastewise and get an answer”

When Chen and his colleagues got started with Tastewise their dream was to create a platform that would democratise data so that any company/user could use it and gain the value they needed, helping them get closer to their consumers. Chen explains that everything changed around 2020 with advancements in generative AI. Before this, Tastewise’s capability did not allow an average, non-tech-savvy user to immediately extract value from the platform. “You had to be comfortable with data, and being able to extract not just the insight, but even acting upon that was not always so easy,” says Chen.

This all changed when Tatsewise launched a Generative AI chatbot, TasteGPT in March of 2023. “Basically, what we did was blend the Gen AI into every part of our platform. So today, you can log into Tastewise and you can use TasteGPT, which is a very natural language chatbot.

“We democratised food and beverage-related data, [...] that means that even my mom can log on to Tastewise and get an answer.”

Chen explains that before Tastewise and TasteGPT, companies often faced a long, and not always easy, process to find answers to their questions. TasteGPT changed this. For example, one company might want to know, “what are the rising diets for Gen Z in the US, is it the same nationwide, or are there any differences regionally? Or what are the fastest trending vegetables people add to their pizza?”. Ask these questions to the GPT chatbot, and you will get an immediate response, Chen says.

“More than that. We don't just give you the data and information, we also take you all the way to implementation and execution. If you need to create a deck for your salesperson who is going to win shelf space, it [TasteGPT] will do that automatically for you.”

Chen uses the example of a company wanting to innovate with a berry and oatmeal yoghurt product. TasteGPT will use the various data sources on the Tastewise platform to create a product concept for the company and help it build a whole story around it.

“It will give you content, it will tell you all [the] consumers [that] are already using similar products in this context or [using] some parts of the product. Because it's a nonexisting product or a novel product, it will then also help you create the whole deck and value proposition.”