Hungryroot

Getting our customers' groceries from the fields to their doorstep takes hard work. On a recent visit to one of our friends and partners, Riverfresh Farms, we got a behind-the-scenes peek at how it all happens—and left more grateful and inspired than ever. Take a look!

This week, Ben McKean, our Founder and CEO, was honored by Goldman Sachs as one of the Most Exceptional Entrepreneurs of 2024, at its Builders and Innovation Summit in Healdsburg, California. Ben joined other entrepreneurs as well as esteemed speakers such as David Solomon, Fred Smith, Candace Parker and more for a two-day event sharing insights on topics ranging from values leadership to the geo-political landscape. The team at Hungryroot is grateful to Goldman Sachs for the recognition. #gsinnovators

161,312. That's how many meals are being donated to local food banks through our in partnership with Mighty Spark.   This summer, we pitched in to support Mighty Spark's mission to donate 1 million meals to Feeding America each year. Every Mighty Spark purchase made through Hungryroot from May through July generated a meal donation. And we're thrilled to share that thanks to this partnership, 161,312 meals will be distributed this fall to food banks that are currently in the greatest need of funds and meals.   Thank you to Mighty Spark for including us in this amazing initiative—we're proud to be part of it!

Prediction: In 2025, you’ll notice this change in how you engage with AI: Today, interactions with AI are primarily human-driven. While AI may answer your search queries, the interaction is initiated by a human. While virtual assistants can respond to voice commands, it’s still the human who initiates the conversation. While AI may recommend products to buy, it’s still the human who adds the items to their cart. This may feel like the desired behavior — do you really want AI incessantly bugging you? — and there are legitimate ethical and privacy concerns with too much AI autonomy. But there are several use cases where we will want AI to just take care of something for us, and we’re nearing that being more commonly the case. Take automated task management, for example. An AI-powered task manager would not only suggest tasks based on your priorities and deadlines, but also proactively schedule and complete tasks on your behalf. For example, it might book meetings, send emails, or even pay bills based on your preferences and habits. Or take what we are building at Hungryroot. Our AI proactively fills your shopping cart based on your predicted needs and delivers them to you without any action required (in fact, 20% of our customers don’t even view SmartCart's selections before purchasing them). As AI continues to evolve, we can expect to see a shift from human-driven to AI-driven interactions. Are you ready for a future in which AI takes the lead? 

What if your idea of healthy eating is different than what your body actually craves? SmartCart, the new AI system we introduced last week at Hungryroot, knows that people’s tastes are more complex than they think. You may believe you only like meat and potatoes, but if you enjoyed that Loaded Carnitas Sweet Potato dish that SmartCart introduced you to, it can recommend other dishes you’re likely to enjoy that you’d be unlikely to choose for yourself. That’s part of the reason Hungryroot customers triple the variety in their diet after joining the service. This also means SmartCart picks up on surprising insights about American eating patterns. For example, customers who tell us they prefer low sodium foods end up buying higher sodium foods than everyone else. When customers explicitly tell us they prefer low sodium foods, as you would expect, SmartCart prioritizes low sodium foods. But these customers tend to remove those recommendations and replace them with higher sodium foods. Why? One potential reason is that people who crave high sodium foods are more likely to have been told that they should eat less sodium. So when they’re asked if they prefer low sodium foods, they say yes, but when they select for themselves, they gravitate toward high sodium foods. This is just one example of the opportunity that SmartCart presents in helping guide people to the foods that will help them feel their best. For these types of customers, SmartCart can identify the low sodium foods that are most attractive to people who crave high sodium meals. Or it can slowly adjust the sodium levels of your food down over time, helping you ease into a new way of eating. There’s a ton of opportunity for SmartCart to help people feel better each day by making it easier to eat healthy. We truly believe we are just getting started. Does your idea of healthy eating differ from what you crave? I’d love to hear from you in the comments. #ai #healthtech #personalizednutrition 

This week marks a major milestone for AI’s ability to improve human health and wellness. Here’s a sneak peek at SmartCart™ 🎉 Food is the foundation of health, and since 2019, Hungryroot has been at the forefront of AI innovation in the food industry. Our original algorithm, known internally as "box fill logic” (or BFL), leveraged an operations research algorithm to automatically fill customers' grocery carts based on their preferences. While two thirds of what our customers bought was chosen by BFL — meaning it underpinned our entire value proposition — it was limited in its ability to dynamically scale and evolve over time. About a year ago, we started developing a new system entirely from scratch by our incredibly talented and experienced in-house team of data scientists, ML and OR engineers. The result is SmartCart™, a first-of-its-kind AI system. It’s comprised of ten machine learning models that integrate into an operations research algorithm and analyzes millions of data points to recommend groceries, recipes, and supplements to help you and your family live healthier, more joyful lives. It’s like having your own personal assistant for healthy living. Each model serves a unique and specific purpose. For example, one of the models, which is built using an approach similar to ChatGPT, analyzes your recent orders, cross-references data from other users, and applies machine learning to suggest the ideal characteristics of your next order — low sodium, quick-to-prepare meals, or high protein, on-the-go snacks, for example. Given the factors that matter most vary by person and evolve over time, there’s a model that applies a weight to each of the other nine models, all designed to optimize customer satisfaction. This flexible, customer-centric approach replaces the rigidity of traditional rule-based systems and is cutting-edge in the field of AI. (Hungryroot has already been granted several patents on the system.) SmartCart™ is already showing results. Customers who use SmartCart™ order twice as often on Hungryroot as customers who shop on their own, demonstrating its ability to help people achieve their health objectives, save time, and discover new foods that bring them joy. 90% of SmartCart™ customers report progress in their health goals since joining the service, and on average, they save 3 hours each week on meal planning, shopping, and cooking. They also discover 3 times more food than shopping on their own. SmartCart™ is a paradigm shift that will require multiple iterations, but I’m so incredibly proud of the team, and I can’t wait to see what our customers think.

Thrilled to have had Ben McKean, Founder and CEO of Hungryroot, on our CEOs You Should Know podcast. In this episode, Ben shares how his team is revolutionizing the way we shop and eat. Learn how Hungryroot is redefining how we approach eating and shopping for food.

The best career advice I’ve received was from Jeff Holden who told me to “never fail silently.” This has stuck with me because it speaks to being ambitious and taking risks, but also working proactively and transparently to minimize risk, two of our core values at Hungryroot. At times we will fail, and that’s okay. We are pursuing an ambitious mission — to lead the grocery industry in using AI to help people live better, healthier lives. But failing silently is not okay. It means not recognizing or clearly communicating when something is not going well. We will face challenges, and as long as we are proactive and transparent, we can work as a team to solve them. We recently recognized this ethos — as well as their extraordinary contributions — in several key promotions at Hungryroot: Dwayne Searwar to SVP Operations, Phil Jeffs to SVP Product, Antoine Atallah to SVP Data Science, Benjamin Zhang to Director Engineering, Wally Chang to Director Engineering, Mayank Gupta to Sr Software Engineer, Sergio Lopez Lazzarini to QA Automation Engineer, Jingchen Cao to Sr Operations Research Engineer, Xiaowei Bao to Principle Operations Research Engineer, Kevin Huang to Sr Data Engineer, Sanat Malhotra to Mgr Food Analytics, Katie Tully to Creative Director, Nicholas Faiello to Sr Photographer, Allison St. Rock to Marketing Mgr, Erin Carroll to Marketing Mgr, Isaac Mccracken to Sr Mgr Transportation, Laura Kershaw to Sr Packaging Engineer, Beth Goodman to Inventory Associate, Jesse Lankford to Lead Product Designer, and Morgan Jenswold to Employee Experience Lead. Congratulations to all of you. I’m fortunate to work with each of you, and grateful that you ambitiously drive our mission forward each day while ensuring that we never fail silently. 

What’s AI hype versus reality? At the Citi GenAI Conference yesterday, we discussed the current state of AI, its promise and limitations, and investor and consumer perceptions. Here are some takeaways: When people talk about AI today, they tend to focus on generative AI, which we’ve become familiar with through tools like ChatGPT. In fact, Citi’s conference wasn’t the AI Conference, it was the GenAI Conference. That’s because generative AI is where major innovations have occurred recently, plus it’s relatable — it can talk to us, create art, write code, and more. At Hungryroot, our sequence predictor system, which we built from scratch to predict what you’ll want next based on your recent orders is generative AI, but we also have a classical AI recommender system and other non-gen AI models that are equally transformative. Over the next couple of years, we believe the focus will return to the opportunity in AI more broadly, with less emphasis on the large language models associated with generative AI. Relatedly, people tend to think about AI as having three layers: infrastructure (NVIDIA, Amazon Web Services (AWS), etc.), models (OpenAI, Google, etc.), and applications (Netflix, Glean, etc.). There’s a lot of debate about where the most value will accrue. Some people think it will look like an hourglass (most of the value to the infrastructure and application layers); others think it’ll be a pear (value to the infrastructure and models). At Hungryroot, we’re most bullish about the application layer and how innovations can disrupt entire industries. At the application layer, it’s clear the focus is currently on the enterprise. At the Citi conference, Hungryroot was the only consumer company that presented; the other 20+ presenting companies sell their AI solutions to businesses. Alongside Sanjay Dhawan, Bob van Luijt and Vishal Marria, I spoke about vertical applications of AI, specifically how it’s disrupting the grocery industry, creating entirely new customer experiences that are more convenient and tailored to personal goals. Lastly, it’s clear leading in AI requires constant innovation and growth since the technology is advancing so rapidly. Our goal at Hungryroot is to lead the grocery industry in using AI to help people live better, healthier lives. We were also recently named one of the top companies in the world using #AI to combat food waste. The opportunity is huge, and we are inspired each day by the challenge. Thank you to Elinor L. Hoover, Jan Metzger and the Citi team for including us in the conference.

Hungryroot is proud to be listed among the top 20 companies in the world using AI to fight food waste. 🌎