Every Friday, Alon Chen’s mother faced a tiny forecasting problem. Her family gathered for Shabbat dinner, but the preferences around the table kept moving. One week brought vegetarians. Another brought keto. Chicken worked until it did not; rice was safe until it was suddenly off the list. The updates arrived in the family WhatsApp group, a stream of new constraints attached to an old ritual. Chen watched his mother try to make one meal hold everyone together and saw something larger hiding inside the chat.
Food companies had the same problem, only with more tables and a much slower group thread. Consumers could change what they cooked, ordered and shared in days. A product team might still need months to run surveys, interpret a report and carry a concept through development. By the time a package reached a shelf, the appetite that justified it could have wandered elsewhere. The family dinner gave Chen a clean way to describe the mismatch: preference moves continuously; industrial knowledge arrives in batches.
That observation became Tastewise, the company Chen founded with technologist Eyal Gaon in 2017. Its early premise was to read the digital traces people leave around food, including menus, recipes, reviews, grocery products and public social conversation, then turn those traces into a living picture of demand. The platform has since expanded from trend analysis into generative and agentic software intended to help marketing, sales and innovation teams do the work that follows an insight.
“My mom makes a great Shabbat dinner, but she started asking us for our dietary needs for that week.”Alon Chen on the origin of Tastewise
A career spent translating
Chen’s route to food technology began well before he had a food company. Born and raised in Israel, he started programming at 12. At 15, he was building and selling computers. Those details can sound like the standard preface to a technical founder story, but Chen adds a useful complication: he can write code, yet he prefers working with people. His career makes more sense as a series of translation jobs, moving between a system and the humans expected to use it.
His first substantial technology job was at Google. Chen joined in Dublin in 2007, initially looking after Hebrew search quality, then moved into marketing. He helped create the program that became Google Partners, taking it across 27 markets and onboarding 60,000 partners. He later served as Google’s chief marketing officer for Israel and Greece and managed the company’s relationship with the World Economic Forum. The work sat at the seam between a technical product, a channel of outside businesses and the local conditions that made the program useful.
Between Google and Tastewise, Chen served as chief business officer at Voyager Labs, where he developed an AI marketing and e-commerce business. The sequence matters. Search taught him to treat messy behavior as a signal. Partner programs taught him that technology spreads through incentives and workflows, not through novelty alone. Marketing leadership taught him to make an unfamiliar system legible. Voyager put him near applied AI before the current wave made the initials unavoidable.
His formal education is similarly mixed: economics at Tel Aviv University, an MBA at University College Dublin and a master’s in law at Bar-Ilan University. Code, markets, institutions and people all appear in the same toolkit. Tastewise would need each one because food is less like a clean software category than a noisy civic square. A word such as “bowl” can describe a format, an occasion, a cuisine signal, an ingredient carrier or simply the object holding lunch.
The weak-signal business
Traditional food research asks a deliberate question of a selected group. Tastewise watches what a much larger public is already doing. The distinction changes both the speed and the texture of the answer. A survey can tell a brand what respondents say they might buy. A menu can show what operators have decided to sell. A recipe can reveal how an ingredient is being used at home. Reviews and public posts add language, mood and occasion. None is complete by itself; together they can show a pattern while it is still forming.
By 2025, Chen described Tastewise’s data universe in unusually large units: tens of billions of social posts, a trillion online recipes, millions of restaurants and hundreds of thousands of grocery items. The numbers provide breadth, but Chen is more revealing when he talks about the limits. “Food is deeply personal and culturally nuanced,” he has said. A signal in one state may mean something different in another. Local flavor preferences, language and context can bend the pattern. The company pairs machine learning with food expertise because a technically valid cluster can still be culturally useless.
A bigger dataset does not remove the need for judgment. It changes where judgment is applied: toward context, interpretation and the choice of what to do next.
This is the tension inside Chen’s project. Tastewise promises scale, but the thing it is scaling is specificity. It groups comparable products, links the same restaurant across delivery services, segments audiences by shared behavior and looks for movement across regions. The ideal result is not a universal answer about what people eat. It is a narrower answer about what a particular group wants, in a particular place, for a particular occasion, while there is still time to respond.
After the answer comes the work
Tastewise’s product history tracks a broader change in enterprise AI. The company began with analytics and predictive intelligence. In 2023 it introduced TasteGPT, a generative interface that could turn the company’s food data into concepts and answers. By 2025 and 2026, Chen was talking less about a clever place to ask questions and more about systems that could participate in execution: preparing campaign material, responding to a customer brief, finding a product opportunity or supporting a sales conversation.
“Intelligence is nothing without action.”Chen’s shorthand for the product shift
The shift sounds subtle, but it changes the value proposition. A dashboard waits for a person to visit, interpret and carry its findings elsewhere. An agent can sit closer to the task. Chen’s 2026 phrasing is direct: “There are no more tools. It’s about implementing agentic systems and helping your team succeed.” He is not arguing that human expertise disappears. His public explanations repeatedly put domain experts and human supervision inside the loop. The ambition is to compress the empty time between knowing and doing.
Investors placed a material bet on that ambition in June 2025. Tastewise announced a $50 million Series B led by TELUS Global Ventures, with participation from Duo Partners, PeakBridge, Disruptive AI and PICO. The company said the round brought its total funding to $72 million and would support expansion across North America, Europe and Asia-Pacific, plus deeper connections into the software food teams already use. The round also marked seven years of work with Gaon, Chen’s co-founder and technical counterpart.
Culture is also a system
Chen’s interest in how systems include or exclude people predates Tastewise. During his first year at Google in Dublin, he was startled that some colleagues worried openness about their partners could damage their careers. He pushed Google to sponsor the local Pride parade. It was a small event then, without an elaborate float, but the visible corporate support mattered to him because it changed what co-workers felt able to say. He later called that outcome the real power of visibility.
As a founder, he has connected that experience to the kind of workplace he wants Tastewise to be. His public account emphasizes representation and the practical value of a mixed team, not just the symbolism. He has also been involved in LGBTQ human-rights advocacy and served on a Tel Aviv municipal advisory board. The through-line is consistent with the product philosophy: people should not have to flatten their differences to fit a system. The system should get better at reading the room.
This makes Chen’s favorite operating themes feel less mechanical than they first appear. He talks about speed, precision and execution, but also about local nuance and human judgment. He learned to code early, yet says he prefers people. He built tools for scale, yet returns to the family table as the clearest explanation of the company. The apparent contradictions are really constraints, and Chen seems most comfortable when several constraints must be held at once.
The table keeps moving
Tastewise now sits at an interesting point in its own forecast. General-purpose AI can summarize a report or draft a campaign in seconds. That makes the surrounding context more valuable: the domain history, cleaned entities, regional distinctions and feedback that tell a system whether its answer belongs in this category at all. Chen’s bet is that food companies will not win by owning the most generic intelligence. They will win by connecting specialized intelligence to the moments where products, menus and messages are actually made.
The family WhatsApp thread remains a useful test. His mother did not need a grand theory of food. She needed a current picture of the people arriving Friday night, then a practical way to serve them. Scale that scene across markets and the data becomes enormous, but the standard of usefulness stays intimate. Did the system notice the change? Did it preserve the context? Did it help someone make a better decision before everyone sat down?
Chen has spent his career moving between machines and markets, global programs and local behavior, prediction and use. Tastewise is the place where those translations converge. The company may count signals by the trillion, but its founding question can still fit inside one message to the family group: what will everyone want this week?